integrative-freight-demand-management-in-new-york

EXECUTIVE SUMMARY

Project Background

Since its inception, this project has received robust support from the industry, rooted in the research conducted for the New York State Department of Transportation (NYSDOT) Initiated in early 2002 at the request of the New York City Chapter of the Council of Supply Chain Management Professionals, the original project titled "Potential for Off-Peak Freight Deliveries to Congested Urban Areas" aimed to explore methods for promoting off-hour deliveries in New York City In response, NYSDOT issued a Request for Proposals in December 2002, leading to Rensselaer Polytechnic Institute being selected as the lead contractor The primary focus of this project was Manhattan, with specific objectives geared towards optimizing freight delivery times.

 ―Define the set of policies and programs that would induce a shift of deliveries to off-peak hours (referred to here as off-peak delivery initiatives).‖

 ―Quantify stakeholders‘ costs and benefits associated with off-peak deliveries initiatives.‖

 ―Perform an economic analysis of the expansion of hours during which pick-ups and deliveries are made to commercial areas.‖

 ―Quantify extra costs to stakeholders so that compensation schemes could be implemented, should off-peak deliveries be found to be economically beneficial to Society at large.‖

In May 2005, the Southwest Brooklyn Industrial Development Corporation (SWBIDC) learned about a project and requested the New York State Department of Transportation (NYSDOT) to include Brooklyn in the study NYSDOT accepted the request, leading to the addition of a Brooklyn-focused second phase to the project The final report for both phases was published on December 8, 2006 (Holguín-Veras, 2006).

The project titled "Integrative Freight Demand Management for the New York City Metropolitan Area" was initiated in March 2007 with funding from the U.S Department of Transportation's Commercial Remote Sensing and Spatial Information Technology Applications Program This funding followed a proposal from a consortium that included Rensselaer Polytechnic Institute, Rutgers University, the Rudin Center for Transportation Policy and Management at NYU-Wagner, and ALK Technologies Inc Initially designed to include a large pilot test, the project's scope was adjusted to focus more on system design, resulting in a transition to a "small scale deployment."

The main charge of the project could be summarized as follows:

The project aims to create a self-sustaining urban freight traffic management system for the New York City metro area, utilizing advanced remote sensing technology, innovative freight demand management, traffic simulation, and policy integration.

The innovative approach merges time-of-day pricing strategies with tax incentives for recipients who agree to off-peak deliveries, alongside GPS-based traffic monitoring This combination aims to encourage a shift in truck traffic to non-peak hours, enhancing revenue generation and reducing congestion.

The work started on July 1, 2007 The total funding provided by the USDOT was about $1.2 million The project partners provided $0.64 million in matching funds

In terms of goals and objectives:

The proposed concept aims to significantly shift truck traffic to off-peak hours, with preliminary estimates indicating a potential 20% reduction in local daytime truck traffic for certain industry segments This shift is expected to enhance traffic congestion and environmental conditions, while also boosting New York City's competitiveness through tax deductions for local businesses, increased productivity from improved traffic flow, and substantial reductions in parking fines, which often surpass typical amounts.

$1,000 per truck per month) Once the concept has been designed and developed, it will be demonstrated in a small scale field deployment test.‖

The project emphasizes urban deliveries, which involve transporting cargo to city locations, as they constitute over 80% of freight traffic in urban areas This focus is crucial for freight demand management programs aimed at alleviating congestion Other segments, such as external-external flows that traverse urban areas, are not addressed in this discussion.

Urban deliveries differ significantly from other freight segments, primarily due to their longer delivery tours, averaging 5.5 stops in New York City, which commence and conclude at a central base Additionally, when cordon pricing is in effect, these tours incur a toll at the urban area's entrance, making the toll surcharge a fixed cost Furthermore, shipments are typically smaller, often requiring the use of smaller vehicles for delivery These unique characteristics highlight the distinct operational challenges associated with urban deliveries in the freight industry.

The findings and methodologies presented in this study are specifically applicable to urban delivery scenarios Additional research is necessary to evaluate their relevance and effectiveness in other freight operation contexts.

Methodology

1.2.1 Remote sensing and pilot test

The pilot test involved four prominent industrial partners interested in off-hour deliveries: Foot Locker collaborating with New Deal Logistics, Sysco alongside a selection of its customers, and Whole Foods Market working with its vendors Notably, all partners are leaders in their respective industries.

In a significant pilot test, 25 receivers and eight carriers/vendors adjusted their distribution chains, particularly transportation and receiving operations, to off-hours for at least a month Each industrial partner group ran their tests independently, commencing when ready, without interactions between groups The commitment from these partners was substantial, involving high-level executives and entire logistics teams in numerous conference calls to prepare for the pilot Although carriers received a token payment of $3,000 as appreciation, this amount barely covered their staff time, highlighting the industry's strong support for the initiative Participation dates are noted alongside company names.

 Group 1: Foot Locker and New Deal Logistics (October 2-November 14, 2009):

Eight Foot Locker stores and New Deal Logistics participated

 Group 2: Sysco and a sample of its customers (December 21-January 23, 2010):

Thirteen stores successfully completed the pilot test, while five others participated partially but withdrew for reasons unrelated to the project Additionally, three stores agreed to participate but did not place orders with the vendor during the test, with the reasons for this remaining unclear.

 Group 3: Whole Foods Market and its vendors (December 28-January 31, 2010):

The study involved four Whole Foods Market locations that were exempt from night delivery restrictions, along with six associated vendors However, two other Whole Foods locations in Manhattan were unable to join the initiative; one faced lease restrictions while the other dealt with neighborhood regulations prohibiting overnight deliveries.

Participants in the pilot test received a financial incentive of $2,000 to encourage successful involvement, which exceeded the incentives considered during the research due to the need for a long-term commitment to off-hours This amount was intended to offset the initial setup costs of transitioning to off-hours and reverting back to regular hours after the pilot Additionally, carriers participating in the test were offered $300 per truck to cover their setup costs, reflecting their potential benefits from off-hour deliveries, allowing for a smaller incentive compared to that of the receivers.

1.2.2 Approach to reducing peak deliveries

Recent analyses from the Port Authority of New York and New Jersey's Time of Day Pricing Initiative and the NYSDOT's Off-Peak Freight Deliveries project reveal significant insights that challenge previous assumptions about urban freight delivery The findings indicate that carriers have limited ability to change delivery times without the agreement of receivers, who typically prefer regular-hour deliveries to utilize existing staff and avoid additional costs Furthermore, the study suggests that cordon tolls are unlikely to encourage off-hour deliveries, as most urban freight carriers cannot transfer these toll costs to customers, which diminishes the effectiveness of price signals Notably, only about 9% of carriers reported being able to pass on toll costs, and approximately 70% cited "customer requirements" as the primary reason for their reluctance to alter delivery schedules Ultimately, this highlights that receivers are the pivotal decision-makers in the delivery process.

Further analyses by Holguín-Veras (2008) indicate that carriers struggle to pass on cordon time-of-day tolls to customers due to a competitive market where delivery rates align with marginal costs As the cordon toll is a fixed cost unrelated to output, it typically does not factor into pricing Empirical evidence shows that only carriers with market power—such as those transporting stone, wood, food, electronics, and beverages—can effectively pass on toll costs This suggests that pricing strategies centered on carriers are less effective, as receivers lack incentives to alter their behavior To address this issue, a new policy paradigm is necessary, focusing on cargo receivers who ultimately influence carriers' ability to adjust delivery times to off-peak hours.

This project aims to shift truck traffic to off-hours by encouraging receivers to accept off-hour deliveries through incentives or by promoting unassisted deliveries that don't require staff Carriers benefit from off-hour deliveries, as they are 20-30% cheaper than regular hours If a sufficient number of receivers agree to off-hour deliveries, carriers can save money by switching entire distribution routes to these times.

Splitting a regular-hour route into two separate routes—one for regular-hour customers and another for off-hours—is unlikely to be advantageous for carriers, as the expenses associated with the additional route would outweigh any operational cost savings.

Inducing receivers to accept off-hour deliveries will lead to the following chain of events:

 The barrier that prevents many carriers from doing off-hour deliveries will be removed

 A significant number of carriers will switch to the off-hours

 Congestion will be reduced, and environmental conditions will improve

 The competitiveness of the urban area will increase as business activities will be more productive and efficient

1.2.3 Behavioral/economic research supporting the concept

Estimating the potential participation in off-hour deliveries is crucial for understanding the economic benefits, including reduced travel times and enhanced environmental conditions This estimation requires a combination of behavioral research to gauge the freight industry's response to various policies and freight trip generation analyses to assess the number of deliveries likely to shift to off-hours Despite employing advanced methodologies, some uncertainty remains in the estimates; therefore, results are presented as ranges whenever possible This section outlines the methodology used and highlights the key findings.

The project’s behavioral research has greatly improved freight transportation modeling and freight behavior analysis, enabling the transportation community to better understand and anticipate the freight industry's reactions to different policies Key outcomes of this research include the creation of a Behavioral Micro-Simulation (BMS) developed by Silas and Holguín-Veras.

In recent studies, significant advancements have been made in freight transportation research, including an approximation model for estimating participation in off-hour deliveries (Holguín-Veras, 2010), the development of an analytical model addressing the limitations of freight road pricing and advocating for comprehensive carrier-receiver policies (Holguín-Veras, 2008; 2009), and the practical application of these innovations to the New York City context.

The behavioral research conducted has led to insight of great practical and theoretical significance More specifically, the research demonstrated on the basis of theory and empirical data that:

 Receiver participation in off-hour deliveries increases with the amount of the incentive provided (though there are industry segments that are more sensitive than others)

Off-hour deliveries can save businesses approximately 30% compared to standard delivery times, as confirmed by cost estimates from the team and input from industry partners.

Delivering during off-hours not only reduces operational costs but also significantly minimizes parking fines, which tend to accumulate during regular hours This financial benefit is crucial, as average parking fines incurred during peak times can be substantial.

$500 and $1,000 per truck per month (Holguín-Veras, 2006) Moreover, since parking fines are not a valid business expense—they are a violation of traffic law—the businesses cannot deduct them from taxes

Carriers with shorter delivery tours are more likely to embrace off-hour deliveries, as they face fewer stops and can more easily secure customer agreement for this option The profitability of off-hour deliveries hinges on a significant portion of customers accepting the arrangement; otherwise, carriers may find it unfeasible to make multiple trips for both regular and off-hour deliveries Thus, those carriers with fewer delivery stops stand to gain the most from implementing off-hour delivery schedules.

Results from Base Case Data and Pilot Test

This section presents the pilot test results, comparing them to base case conditions through productivity performance measures and feedback from participating companies The analysis emphasizes travel speeds from the depot to the first customer in Manhattan and between customers, as well as service times during deliveries The team distinguished between the longer initial trip and the shorter, more frequent customer-to-customer trips, highlighting the unique challenges of urban driving conditions By focusing on these travel speeds, the analysis captures the cumulative effects of traffic delays Additionally, examining service times sheds light on delivery-related delays The findings were validated through discussions with industrial partners Due to insufficient data from individual groups for statistically significant results, the datasets were combined to create more reliable estimates.

The speeds reported in this analysis are space mean estimates, calculated by dividing the distance from a point of origin to a destination by the travel time, factoring in interruptions from traffic signals, pedestrians, and other vehicles Instantaneous speeds are excluded due to their variability and inability to reflect these obstructions Readers should note that data is unavailable between 10 PM and 4 AM, with limited observations during the 4-5 AM period While the dataset comprises approximately 4,000 individual trips, its representativeness of overall truck traffic in the New York City metropolitan area cannot be guaranteed Nonetheless, the data offers valuable insights into the potential impacts of off-hour deliveries.

The results are illustrated using a box and whisker plot, which highlights key percentiles: the 2nd, 25th, 50th (median), 75th, and 98th, along with any outliers Percentiles represent the value below which a specific percentage of observations fall; for example, the 25th percentile indicates that 25% of the data points are below this value In the plot, the 2nd percentile is marked at the lower whisker's tip, the 25th percentile at the lower box tip, the median at the line within the box, the 75th percentile at the top of the box, and the 98th percentile at the upper whisker's tip Outlier values are indicated by crosshatches outside these percentiles.

1.3.1 Customer to customer travel speeds

Customer-to-customer speed data reveals a distinct trend, with speeds peaking during early morning hours and declining throughout the day As shown in Figure 1, speeds reach nearly 8 miles per hour between 5 and 7 AM, while dropping to below 3 miles per hour during daytime hours The limited data from 4 to 5 AM, consisting of only four observations, indicates that travel speeds could potentially be even higher during these early hours.

Off-hour delivery tours can significantly enhance productivity, with a truck traveling ten miles potentially saving 1.25 hours of travel time by operating at an average speed of 8 mph during off-hours compared to just 4 mph during regular hours The economic benefits of switching to off-hour deliveries increase with the length of the tours, making it a more efficient choice for longer routes.

Figure 1: Customer to Customer Space Mean Speeds by Time of Day

The second key performance measure is service time, which refers to the total duration a driver spends at a customer's location This encompasses various activities such as loading and unloading cargo, walking to and from the truck, locating the recipient for the delivery, waiting for the necessary approvals and signatures, addressing any issues that may arise, and other related tasks Figure 2 illustrates the estimates generated by the team.

Service times for deliveries exhibit a notable pattern, increasing during daytime hours and decreasing significantly during off-hours In the morning, when most deliveries occur, service times consistently surpass one hour, peaking at a median of 1.8 hours between 10 AM and noon Conversely, during nighttime hours, median service times drop to approximately half an hour Although the representativeness of these figures for the entire industry remains uncertain, they suggest that carriers could potentially save up to 1.3 hours per delivery by shifting operations from morning to night.

The team engaged with industrial partners to validate their findings, which the partners affirmed reflect the realities of operating in New York City, highlighting them as part of the "cost of doing business." During daytime hours, drivers often face challenges such as parking 2-3 blocks away from customer locations, waiting for loading docks, experiencing delays accessing elevators due to other deliveries or building visitors, and relocating trucks to avoid fines However, during off-hours, these issues significantly decrease, allowing drivers to park closer to customers and streamline their operations.

The team inquired about delivery sizes during regular and off-hours, revealing that off-hour shipments are generally larger due to increased productivity, allowing for the transport of more cargo This finding rules out the possibility that longer service times during regular hours are caused by larger shipments Consequently, when accounting for the larger shipment sizes during off-hours, the productivity savings from reduced service times are likely greater than those indicated in Figure 2.

Figure 2: Service Times by Time of Day

Reducing service times in delivery operations can significantly enhance profitability and lower product costs in New York City For instance, a delivery truck that saves 15 minutes on each of six deliveries can reduce its total delivery time by 1.5 hours, translating to a savings of $60 per tour Similarly, a carrier that saves an average of 30 minutes per delivery could save around three hours in total These time savings lead to substantial economic benefits, underscoring the importance of efficiency in the delivery industry.

While the estimates highlight the advantages of off-hour deliveries for participating carriers, it's crucial to remember that they do not account for the benefits of reduced truck traffic during peak hours, which are addressed in the traffic simulation section.

On November 11, 2009, Mr Paul Cox, Vice-President for Global Transportation and Supply Chain at Foot Locker, shared insights on the company's positive experiences with off-hour deliveries, especially in larger volume stores that have dedicated backroom staff.

12 operations As of the project team‘s last communication with Mr Paul Cox, Foot Locker was considering expanding off-hour deliveries to other stores in Manhattan

The team conducted satisfaction surveys for Foot Locker, but communication issues affected the responses from store managers At the time of the survey, managers were unaware that headquarters had secured a financial incentive to cover additional costs, leading them to view off-hour deliveries negatively The average rating for these deliveries was 3.88 on a scale of one to five, indicating a generally unfavorable attitude without the context of the financial incentive Consequently, the survey results did not accurately reflect the project's intentions In later versions of the survey, the team resolved these communication problems.

Mr Rob Twyman, the Regional Vice President of Operations for Whole Foods Market in the Northeast Region, reported positive feedback from stores regarding the recent shift in delivery times for participating vendors, describing the transition as "relatively seamless." Additionally, the project team learned that several vendors involved in the pilot test are now consistently delivering to Whole Foods Market during off-hours.

The project team analyzed satisfaction surveys from twelve Sysco receivers regarding off-hour deliveries, resulting in an average overall impression score of 1.50 on a scale of one to five, where "1" indicates "Very Favorable" and "2" denotes "Favorable." When asked about the likelihood of requesting off-hour deliveries in the future, nine respondents indicated they were "Very Likely," one was "Likely," and two were uncertain Additionally, six participants utilized unassisted deliveries during the pilot, with five of the remaining six expressing interest in future unassisted deliveries, contingent upon resolving liability concerns.

Economic Impacts of a Full Implementation

The economic impacts of fully implementing an off-hour delivery program were assessed by quantifying the total number of truck trips that would transition to off-peak hours.

This chapter outlines the methodology and key findings related to a financial incentive designed to offset additional costs It discusses the application of network models to simulate various scenarios associated with shifting to off-peak hours Additionally, it estimates the economic value of impacts, including travel time savings and environmental benefits.

1.4.1 Quantification of the potential switch to the off-hours

Estimating potential participation in off-hour deliveries involves a two-step process: first, assessing market response to financial incentives, and second, quantifying the number of deliveries generated across different industry segments.

The market response to financial incentives for receivers was estimated using a BMS model (Silas and Holguín-Veras, 2009) and an analytical model (Holguín-Veras, 2010), both of which yielded similar outcomes These models are specifically designed to assess the market's reaction to incentives and various policies.

The article discusses the estimation of how many receivers in simulated delivery tours would opt for off-hours delivery These estimates utilize advanced discrete choice models, which are calibrated using stated preference data collected by the research team, to assess the potential behavior of receivers when presented with incentives.

The team's survey data provides detailed insights into delivery tours, the number of stops, and operational patterns, enabling them to simulate carrier responses effectively By analyzing the financial impact of receivers switching to off-hour deliveries, they calculated the associated cost savings or increases Typically, if the majority of receivers opt for off-hour deliveries, carriers experience reduced costs and are inclined to accommodate these requests Conversely, if only a few receivers choose this option, delivery costs may rise, leading carriers to decline off-hour services.

The team aggregated results across different industry segments by simulating outcomes from randomly generated tours This analysis allowed them to estimate the number and percentage of receivers and carriers willing to perform off-hour deliveries in sectors such as food and retail.

The team aimed to quantify the total number of deliveries across different industry segments to assess the effects of off-hours deliveries on the transportation network.

 Post processed the survey data collected to estimate freight trip generation models (the only models available for New York City conditions)

 Used the models to estimate the total number of deliveries by industry segment and ZIP code for the Manhattan area

 Computed the number and percentage of deliveries that would switch to the off-hours for each ZIP code

 Used these estimates to modify the inputs to the Best Practice Model (BPM), and a mesoscopic traffic simulation model developed by the research team for Midtown and Downtown Manhattan

Manhattan experiences an impressive volume of freight deliveries, with approximately 113,000 deliveries made daily This equates to an average of 0.163 deliveries per employee and 2.798 deliveries per establishment The area's consumer-driven nature is evident, as the Consumer Goods sector accounts for 54.48% of deliveries, while the Food sector, including eating and drinking establishments as well as food stores, represents 24.28%.

The Food and Consumer Goods sectors are identified as the most suitable candidates for off-hour deliveries due to their high delivery volume and willingness to participate The team anticipates that offering financial incentives ranging from $5,000 to $10,000 per year could encourage 7-15% of truck trips to shift to off-peak hours, resulting in an increase of 7,000 to 16,000 truck trips per day.

 Large Traffic Generators (LTGs) generate about 4-8% of the total number of freight deliveries in Manhattan This is a notable result Although there are no hard data to accurately quantify the share of LTGs in the total freight traffic, the team estimated that the

In Manhattan, 88 buildings with their own ZIP codes, including iconic structures like the Empire State Building, account for approximately 4% of the total freight traffic When considering additional locations without unique ZIP codes, such as Grand Central Terminal and the Javits Center, along with other significant establishments that contribute to truck traffic, it's likely that large transportation generators (LTGs) in the area produce between 4% and 8% of the total truck traffic Notably, LTGs typically feature centralized delivery stations, allowing them to receive off-hour deliveries and distribute them to businesses during regular operating hours.

 Unassisted deliveries provide a great alternative to the provision of financial incentives

Unassisted deliveries are innovative systems that eliminate the need for human intervention at the delivery point, enhancing efficiency and security These systems include key deliveries, where drivers use a key provided by the receiver to deposit goods inside a store; double-door setups that allow deliveries to be secured in a designated area; electronic delivery lockers for convenient retrieval by the consignee; and two-stage delivery systems that involve off-hour transport to secure locations for later delivery using eco-friendly vehicles Despite these advancements, there is a lack of behavioral data to gauge industry willingness to adopt unassisted delivery methods The following section details the network modeling and traffic simulation processes involved.

1.4.2 Network modeling and traffic simulation

The team evaluated the traffic effects of various off-hour delivery scenarios using the New York Metropolitan Transportation Council’s Best Practice Model (BPM) and a mesoscopic traffic simulation model tailored to an extracted network in Manhattan This comprehensive analysis involved multiple activities to assess the potential impacts on traffic flow.

The New York Best Practice Model was enhanced by acquiring the latest hourly volume data from transportation agencies in the New York area This updated data was essential for calibrating the model to reflect the most current truck traffic volumes, ensuring accuracy and relevance in transportation planning.

Conclusions and Suggested Next Steps

The project has received widespread acclaim from the freight industry, research community, and various agencies as a groundbreaking initiative worthy of replication and expansion Key highlights include successful meetings with the Industry Advisory Group (IAG) and the Technical Advisory Group (TAG) on December 9, 2009, and the publication of two positive articles in the esteemed Journal of Commerce Additionally, it garnered attention from the Wall Street Journal and was recognized by NYCDOT Commissioner Janette Sadik-Khan for its significant potential impact in New York City Ultimately, the project has demonstrated its effectiveness in shifting urban deliveries to off-hours and has garnered strong support from industry stakeholders.

(3) would bring about substantial reductions in congestion and environmental pollution thus increasing quality of life; and (4) would increase the competitiveness of the urban economy The fact that this is a

22 win-win concept that benefits all the participants in urban deliveries provides a unique opportunity for expansion and full implementation The analyses conducted by the team indicate that:

Financial incentives for receivers can effectively encourage carriers to shift deliveries to off-peak hours By compensating receivers for the additional costs associated with off-hour deliveries, they are motivated to make the switch Meanwhile, carriers benefit from reduced delivery costs and fewer parking fines during these times, making the transition appealing for them as well Analyses suggest that, depending on the industry segment and the incentives offered, this shift could reduce truck traffic by 10% to 20%.

Traffic simulations reveal that shifting truck traffic to off-peak hours can yield significant economic advantages Estimates suggest that an optimal financial incentive of just over $10,000 annually would be embraced by approximately 7,600 establishments, resulting in a total cost of $76 million The anticipated economic benefits from this shift range from $83 million to $129 million, influenced by the value of travel time considered in the analysis Notably, beyond the $10,000 incentive, the additional benefits diminish while the associated costs continue to rise.

GPS data reveals that trucks operating during off-peak hours average speeds of 8 miles per hour, compared to less than 3 miles per hour during peak times Consequently, a truck covering a distance of 10 miles can save approximately 1.25 hours per delivery route when scheduled during off-peak hours.

During off-hours, service times experience significant reductions, averaging just half an hour per customer In contrast, regular hours often see service times exceeding 1.5 hours due to factors such as long walks from parking, elevator congestion, and delays in customer check-ins for deliveries With these obstacles largely eliminated during off-peak times, the overall savings in service time are considerable, likely surpassing the savings in travel time for delivery trucks serving multiple customers in a single tour.

Despite the promising concept and positive outcomes from the project, several critical questions must be addressed before full implementation Key concerns include the noise impact on nearby communities, the statistical validity of results from the small pilot test, the potential effectiveness of targeted programs for large traffic generators, the encouragement of unassisted off-hour deliveries, and the need for inter-agency coordination and policy development Addressing these questions is essential for ensuring a successful and sustainable implementation.

 Noise impacts were not assessed during the project Although no community complaints were received during the execution of the small pilot test, it is natural to expect that

Noise impacts are a significant concern for 23 community members, making it essential to evaluate these effects and develop suitable mitigation strategies if noise poses a potential barrier to implementation The primary objective is to protect local communities from any negative consequences.

The small size of the pilot test conducted does not yield statistically representative results, highlighting the need for a larger scale study to effectively integrate remote sensing into a functional prototype and evaluate the benefits of off-hour deliveries Both team members and the USDOT have acknowledged this limitation Although an expansion of the pilot was considered at the end of 2008, it was ultimately rejected due to the unfavorable economic climate at the time However, with recent improvements in economic conditions and the stability of financial markets, there is now a significant opportunity to expand the pilot test and advance this innovative research.

 About 4-8% of all deliveries to New York City are generated by Large Traffic Generators

Inducing Light Trucks and Goods (LTGs) to conduct off-hour deliveries could significantly alleviate traffic congestion in New York City Given the relatively small number of LTGs, ranging from 90 to 500 depending on the definition, the coordination required for this shift is minimal compared to the potential benefits Therefore, the City of New York could play a pivotal role in encouraging LTG owners to adopt off-hour delivery schedules as part of its sustainability initiatives.

Unassisted off-hour deliveries can significantly enhance sustainability strategies by providing financial benefits at a lower cost Addressing liability concerns through public sector programs is crucial for encouraging businesses to adopt this practice As more companies become familiar with unassisted off-hour deliveries, participation is likely to grow Notably, 80% of receivers involved in a recent study expressed willingness to engage in unassisted off-hour deliveries if liability issues were addressed.

Effective inter-agency coordination is essential for successfully implementing off-hour delivery strategies, which significantly impact economic development, environmental sustainability, and energy conservation Engaging relevant agencies in the creation of a unified off-hour delivery approach will result in strong policies and seamless execution Promoting off-hour deliveries at major traffic generators and maximizing unassisted deliveries is crucial, as it reduces costs and logistical challenges for receivers.

Off-hour deliveries are cost-effective, requiring fewer incentives than traditional programs, yet significant questions persist regarding their implementation at major traffic generators Key concerns include the integration of remote sensing for compliance, liability issues, cost-benefit analysis for participants, and the effectiveness of alternative policies.

Macroscopic and mesoscopic traffic models significantly reduce congestion and enhance environmental conditions, with regional and sub-regional models yielding varying estimates of benefits Integrating these models is vital for accurately assessing dynamic traffic impacts at the facility level While current modeling predominantly addresses short-term effects, understanding long-term network-wide impacts necessitates extensive data collection and analysis An extended pilot test will facilitate better data access and model calibration Utilizing these simulation models is essential for comprehending various scenarios, quantifying their effects, and gaining support from relevant agencies The research team, armed with extensive experience, is well-positioned to expand these simulation studies further.

All of this suggests the steps outlined below:

To enhance off-hour deliveries in New York City, a comprehensive approach is needed that includes behavioral research to identify cost-effective incentive policies, alongside the expansion of traffic simulation models for accurate network representation Collaboration with city and state agencies, such as the New York City Economic Development Corporation and the New York State Energy Research and Development Agency, will be essential for successful policy implementation A publicity campaign should be launched to garner industry support and recruit participants for an expanded pilot test Additionally, research on noise impacts and mitigation strategies, along with community feedback programs, will ensure that local concerns are addressed Finally, establishing compliance verification mechanisms will be crucial for the effective enforcement of these initiatives.

The expanded pilot test and implementation phase will involve rolling out a comprehensive set of incentive policies and recruiting participants A detailed monitoring plan will be designed, utilizing GPS devices to evaluate delivery operations' performance before and after the pilot, along with noise measuring devices to assess environmental impacts Additionally, existing GPS data from NYCDOT will be used to monitor network-wide effects The expanded pilot test will be launched and closely monitored, employing behavioral models to predict participant engagement during implementation Traffic simulation models will be used to analyze the network impacts of both the expanded pilot test and full implementation Community hearings will be organized to gather stakeholder input, leading to a final decision on the implementation process.

INTRODUCTION

Freight Road Pricing and Comprehensive Carrier-Receiver Policies

Road pricing aims to align private driving costs with the social costs of driving, effectively eliminating deadweight losses Research indicates that this approach is a successful transportation demand management strategy, enhancing economic welfare while generating substantial revenue for transportation investments (Sullivan, 2000) Notably, in passenger transportation, the decision maker is also the unit of demand, highlighting the unique dynamics of this market.

From a behavioral perspective, the impact of tolls is directly experienced by those making travel decisions However, urban freight presents a different scenario, as evidence suggests that freight road pricing is often ineffective in shifting truck traffic to off-peak hours This ineffectiveness can be attributed to various factors, including market imperfections and contractual constraints Most notably, delivery times are largely determined by receivers, with 40% of cases set solely by them, 38% jointly with carriers, and only 22% by carriers alone As customers of carriers, receivers play a crucial role in determining delivery schedules, making it essential for them to be willing to operate during off-peak hours for freight road pricing to be effective Ultimately, the success of such pricing strategies hinges on the strength of the price signals from carriers and the receivers' readiness to adapt their operations accordingly.

Examining the empirical evidence on pricing's impact on carrier behavior reveals significant insights A study conducted on the implementation of time-of-day pricing at the Port Authority of New York and New Jersey in 2001 indicated that 20.2% of carriers altered their behavior due to this pricing initiative Interestingly, the carriers' responses were not straightforward; instead, they engaged in complex, multi-dimensional strategies that included increases in productivity, cost transfers, and other adaptive measures.

Recent data reveals that changes in facility usage reflect a complex response rather than a straightforward adjustment Notably, three strategic combinations account for nearly 90% of cases: Productivity increases (42.79%), Changes in facility usage and Cost transfers (27.60%), and a combination of all three factors (19.32%) It's important to note that while some responses, like Productivity increases, primarily affect carriers, others have a more significant impact on receivers.

Changes in facility usage and cost transfers highlight that the response dynamics are influenced by the power balance between carriers and receivers Notably, 69.8% of carriers who maintained their behavior attributed this decision to customer requirements, indicating a strong influence of client demands on operational flexibility.

Only 9.0% of the sample opted to raise shipping charges for receivers, highlighting the fragility of the urban delivery carrier industry due to low entry costs fostering competitiveness Additionally, the average increase in shipping costs was modest, around 15%, as carriers typically distribute toll hikes among their pricing structures.

28 among the multiple customers in a delivery route, about 5.6 deliveries per tour (Holguin-Veras et al.,

The limited price signals from carriers primarily affect a small number of receivers, highlighting the necessity to expand transportation policy to consider the essential role of customers in determining delivery times Consequently, relying solely on freight road pricing is unlikely to effectively encourage a substantial shift of truck traffic to off-peak hours, as the signals reaching receivers are insufficient to motivate them to adjust their operational hours.

The analysis of behavioral data reveals two key insights: first, carriers have limited ability to unilaterally alter delivery times, as receivers prefer regular hours for efficiency and cost reasons; second, cordon tolls are ineffective in encouraging off-hour deliveries since most urban freight segments cannot pass these costs to customers, lacking the necessary price signals for behavioral change Further research indicates that only market segments with sufficient power, such as those transporting stone, wood, food, electronics, and beverages, can pass toll costs effectively Consequently, a new policy paradigm is required, focusing on both carriers and receivers, termed "carrier-receiver" policies, which aim to integrate carrier-centered approaches like freight road pricing with receiver-centered strategies.

A series of studies evaluated various carrier-receiver policies related to off-hour delivery (OHD), highlighting implementation constraints (Holguín-Veras, 2006) and the restaurant sector's potential as a target for off-hour delivery programs (Holguín-Veras et al., 2006b) The research provided a framework for analyzing carrier-receiver interactions and examined behavioral models derived from stated preference data from both parties (Holguín-Veras et al., 2007; 2008) The findings revealed that receivers are willing to adopt off-hour deliveries for financial incentives, carriers are responsive to receiver requests, and only a few industry segments, such as petroleum/coal, wood/lumber, and food carriers, show sensitivity to tolls.

Research shows that carriers favor policies promoting off-hour deliveries, as these are approximately 30% more cost-effective than deliveries made during peak congestion times (Holguín-Veras, 2006) This finding highlights the potential benefits of implementing such policies for products, textiles, and clothing.

Previous Experiences with Off-Hour Deliveries

The concept of implementing Off-Hours Deliveries (OHD) to alleviate urban congestion is not a recent development; it dates back to the Roman era, specifically outlined in Julius Caesar's "Lex Juliana Municipalis," which required goods deliveries in Rome to occur during the evening to minimize congestion Historical accounts indicate that Roman citizens expressed concerns over the increased noise generated during these off-hours, highlighting that noise pollution continues to be a significant challenge for OHD initiatives today.

Recent studies on Off-Hours Delivery (OHD) have revealed mixed findings regarding its effectiveness The first significant study in London (1968) indicated that OHD could be beneficial for large shipments with fewer deliveries (Churchill, 1970) However, a 1979 study by the Organization for Environmental Growth highlighted the need for pilot testing to clarify OHD's impacts, as feedback from various stakeholders was inconclusive (OFEGRO, 1979) Further research in the late 1970s (Noel et al., 1980) found that while delivery companies appreciated the cost and time savings from OHD, security concerns hindered its implementation The Urban Gridlock Study suggested that OHD could moderately alleviate traffic congestion, but the space freed by trucks would likely be filled by passenger vehicles (Grenzeback et al.; Cambridge Systematics, 1988b; a) In Los Angeles, a proposal to ban trucks during peak hours faced strong opposition from businesses due to increased operational costs (Nelson et al., 1991) Additionally, the Port Authority of New York and New Jersey found that trucking firms generally try to avoid peak travel times but expressed skepticism about the effectiveness of peak-hour tolls in reducing congestion (Vilain and Wolfrom, 2000).

Recent pilot tests were conducted to understand how OHD would impact traffic and the environment in Athens, Barcelona, Dublin, and London The Athens study examined land use, delivery requirements

A study by Yannis et al (2006) demonstrated that off-hour deliveries (OHD) could significantly alleviate traffic congestion and environmental pollution by shifting truck traffic to non-traditional hours, resulting in lower emissions of nitrogen oxides, hydrocarbons, and sulfur dioxide Pilot programs in Barcelona and Dublin, which involved reducing the number of deliveries to supermarkets, showed that OHD led to decreased logistical delays, traffic congestion, and energy consumption, while enhancing road safety and providing economic benefits through reduced shipping costs In Wadsworth, London, OHD operations using low-noise equipment and larger trucks for Sainsbury’s resulted in substantial savings of approximately $25,000 annually, alongside increased worker efficiency and positive customer feedback Overall, these initiatives reflect a growing interest in implementing OHD strategies to improve urban logistics.

Project Background

Since its inception, this project has received robust industry support, rooted in the work carried out for the New York State Department of Transportation (NYSDOT) by team members Initiated in early 2002 at the request of the New York City Chapter of the Council of Supply Chain Management Professionals, the original NYSDOT project, titled "Potential for Off-Peak Freight Deliveries to Congested Urban Areas," aimed to explore methods to promote off-peak freight deliveries in New York City In response, NYSDOT recognized the importance of the study and issued a Request for Proposals in December 2002, ultimately selecting Rensselaer Polytechnic Institute as the lead contractor The project's primary focus was on Manhattan, with specific objectives aimed at addressing urban freight congestion.

 ―Define the set of policies and programs that would induce a shift on deliveries to off-peak hours (referred to here as off-peak delivery initiatives).‖

 ―Quantify stakeholders‘ costs and benefits associated with off-peak deliveries initiatives.‖

 ―Perform an economic analysis of the expansion of hours during which pick-ups and deliveries are made to commercial areas.‖

 ―Quantify extra costs to stakeholders so that compensation schemes could be implemented, should off-peak deliveries be found to be economically beneficial to Society at large.‖

In May 2005, the Southwest Brooklyn Industrial Development Corporation (SWBIDC) learned about a project through informal channels and requested the New York State Department of Transportation (NYSDOT) to include Brooklyn in the study NYSDOT complied, adding a Brooklyn-focused second phase to the project, and the final report for both phases was published on December 8, 2006 (Holguín-Veras, 2006).

The project "Integrative Freight Demand Management for the New York City Metropolitan Area," funded by the U.S Department of Transportation in March 2007, was initiated following a proposal from a consortium including Rensselaer Polytechnic Institute, Rutgers University, the Rudin Center for Transportation Policy and Management at NYU-Wagner, and ALK Technologies Inc Initially designed for a large-scale pilot test, the project's scope was adjusted to focus more on system design, resulting in a shift to a "small scale deployment."

The main charge of the project could be summarized as follows:

The project aims to create a self-sustaining urban freight traffic management system for the New York City metro area, utilizing advanced remote sensing technology, innovative freight demand management, and traffic simulation alongside effective policy measures By integrating time-of-day pricing with tax incentives for receivers who accept off-peak deliveries, and employing GPS-based traffic monitoring, the initiative seeks to encourage a shift of truck traffic to non-peak hours, thereby optimizing urban freight logistics.

Funding

The work started in July 1, 2007 The total funding provided by the USDOT was about $1.2 million The project partners provided $0.64 million in matching funds.

Goals and Objectives

The project aimed to investigate the integration of GPS-enabled smartphones and financial incentives for urban delivery recipients, creating a comprehensive freight demand management system that addresses the shortcomings of traditional pricing strategies The key goals and objectives of the project focused on enhancing urban logistics efficiency through innovative technology and incentive-based approaches.

The proposed concept aims to significantly shift truck traffic to off-peak hours, potentially reducing local daytime truck traffic by up to 20% in certain industry segments This shift is expected to alleviate traffic congestion and enhance environmental conditions Additionally, it will boost New York City's competitiveness through tax deductions for local businesses, increased productivity from improved traffic flow, and substantial reductions in parking fines, which often exceed current levels.

$1,000 per truck per month) Once the concept has been designed and developed, it will be demonstrated in a small scale field deployment test.‖

This study focuses on congested urban areas, with New York City serving as a case study, highlighting the significant benefits of OHD in such environments The concepts and methodologies developed in this research are relevant and can be applied to other large metropolitan regions.

The project primarily targets urban deliveries, which constitute over 80% of freight traffic in urban areas This focus is crucial for implementing freight demand management programs designed to alleviate congestion caused by these deliveries Other freight segments, such as external-external flows that traverse urban spaces, are not covered in this discussion.

Urban deliveries differ significantly from other freight segments due to their unique characteristics Typically, these deliveries involve longer tours, averaging 5.5 stops in New York City, that begin and end at a central base Additionally, when cordon pricing is in effect, delivery tours incur a fixed toll cost at the urban area's entrance Furthermore, shipment sizes are generally small, often requiring the use of smaller vehicles for transportation These factors contribute to the distinct nature of urban delivery operations, setting them apart from other freight industry segments.

The findings and methodologies established in this study are specifically applicable to urban delivery scenarios Additional research is necessary to determine their relevance and effectiveness in other freight operation contexts.

References

Cambridge Systematics (1988a) Urban Gridlock Study Technical Report California Department of Transportation

Cambridge Systematics (1988b) Urban Gridlock Study: Summary Report California Department of Transportation

Churchill, J D C (1970) Operation "MoonDrop" : An Experiment in out of Hours Goods Delivery Proceedings of the 3rd Technology Assessment Review, Paris, France

Desau, H (1892) "Inscriptiones Latinae Selectae, No 6085."

Grenzeback, L R., W R Reilly, P O Roberts and J R Stowers "Urban Freeway Gridlock Study: Decreasing the Effects of Large Trucks on Peak-Period Urban Freeway Congestion." Transportation Research Record 1256

Holguín-Veras, J (2006) Potential for Off-Peak Freight Deliveries to Congested Urban Areas NYSDOT http://www.rpi.edu/~holguj2/OPD/OPD_FINAL_REPORT_12-18-06.pdf

Holguín-Veras, J (2008) "Necessary conditions for off-hour deliveries and the effectiveness of urban freight road pricing and alternative financial policies in competitive markets." Transportation Research Part A: Policy and Practice 42(2): 392-413

Holguin-Veras, J., K Ozbay and A de Cerreno (2005) Evaluation Study of the Port Authority of New York and New Jersey's Time of Day Pricing Initiative NJODOT

Holguín-Veras, J., N Pérez, B Cruz and J Polimeni (2006a) "Effectiveness of Financial Incentives for Off-Peak Deliveries to Restaurants in Manhattan, New York." Transportation Research Record: Journal of the Transportation Research Board 1966(-1): 51-59

Holguín-Veras, J., M A Silas and J Polimeni (2006b) "On the Overall Performance of Comprehensive Policies to Manage Truck Traffic in Congested Urban Areas." First International Conference on Funding Transportation Infrastructure

Holguín-Veras et al (2007) conducted a study published in "Networks and Spatial Economics" that examines the effectiveness of joint receiver-carrier policies aimed at increasing truck traffic during off-peak hours The research focuses on the behaviors of receivers and highlights strategies that could optimize freight operations while alleviating congestion during peak times This investigation provides valuable insights for policymakers and logistics managers seeking to improve transportation efficiency and reduce delays.

Holguín-Veras et al (2008) investigated the effectiveness of joint receiver-carrier policies aimed at increasing truck traffic during off-peak hours, revealing insights into carrier behaviors Additionally, Holguín-Veras et al (2006) examined the impacts of time-of-day pricing on freight carriers in congested urban areas, highlighting significant implications for road pricing strategies These studies contribute valuable knowledge to transportation policy, emphasizing the importance of strategic pricing to manage traffic flow and improve urban mobility.

Nelson, C., S Siwek, R Guensler and K Michelson (1991) Managing trucks for air quality Paper presented at 70th Annual Meeting of the Transportation Research Board, Washington, D.C

NICHES (2008) "Innovative Approaches in City Logistics: Inner-City Night Delivery." Retrieved March

26, 2010, from http://www.niches- transport.org/fileadmin/archive/Deliverables/D4.db_5.8_b_PolicyNotes/14683_pn7_night_delivery_o k_low.pdf

Noel, E C., S H Crimmins, N K Myers and P Ross (1980) "A Survey of Off-Hours Delivery." ITE Journal

OFEGRO (1979) Requirements and Specifications for Off-Hour Delivery Organization for Environmental Growth

The final report by E Sullivan (2000) from the California Department of Transportation assesses the effects of the SR 91 Value-Priced Express Lanes This comprehensive study evaluates traffic operations and provides insights into the impacts of value pricing on express lanes For further details, the report can be accessed at the provided link.

Texas Transportation Institute (2009) "Urban Mobility Report." Retrieved July 10, 2009, from http://mobility.tamu.edu/ums/report/

Vilain, P and P Wolfrom (2000) "Value Pricing and Freight Traffic: Issues and Industry Constraints in Shifting from Peak to Off-Peak Movements." Transportation Research Record 1707: 64-72

Yannis, G., J Golias and C Antoniou (2006) "Effects of Urban Delivery Restrictions on Traffic Movements." Transportation Planning and Technology 29: 295-311

INSTITUTIONAL ANALYSIS, COORDINATION, OUTREACH, AND

Description of Institutional Setting

The greater New York metropolitan area, with nearly 20 million residents across three states and numerous municipalities, relies heavily on its freight system, which annually transports approximately 47 million tons of food, 2.8 million tons of clothing, and 70 million tons of gasoline essential for daily life However, the efficiency of this system is challenged by severe traffic congestion and aging infrastructure, along with the complexities arising from multiple jurisdictions involved in surface transportation planning and execution.

The freight movement in the region is hindered by a fragmented governance structure involving multiple public agencies The New York Metropolitan Transportation Council (NYMTC) oversees the creation of a Regional Transportation Plan (RTP) for downstate New York but does not include New Jersey or Connecticut While the RTP outlines a regional vision, the implementation of specific projects falls to various agencies, such as the New York State Department of Transportation (NYSDOT) for highways and the New York City Department of Transportation (NYCDOT) for local systems Additionally, critical infrastructure for goods movement, including ports, bridges, and tunnels, is managed by special agencies like the Port Authority of New York and New Jersey (PANYNJ) and the Metropolitan Transportation Authority (MTA) This complex system is further complicated by the significant freight traffic between New Jersey and New York City, involving multiple state agencies.

The North Jersey Transportation Planning Authority plays a crucial role in freight transportation planning, coordinating with private sector entities responsible for goods delivery across the region These private companies face a complex web of regulations imposed by federal, state, and local governments in the three states This intricate institutional framework highlights the necessity for inter-jurisdictional cooperation, as formal integration is unlikely Therefore, fostering forums for collaboration is vital for implementing effective regional strategies, underscoring the importance of outreach in this project.

Public Outreach and Institutional Analysis

Successful implementation of OHD requires significant input from both industry and public sector stakeholders To facilitate this, the project team created a comprehensive list of potential contacts, leading to the formation of an Industrial Advisory Group (IAG) and a Technical Advisory Group (TAG) Through various communication methods, including phone calls, emails, and in-person meetings, six representatives from the private sector joined the IAG, while seven representatives from public agencies committed to participating in the TAG.

In Task 2 (IC2), the project team conducted interviews with IAG and TAG members to gather insights on delivery time shifts and policy challenges related to the proposed OHD The IAG focused on understanding concerns about changing delivery schedules, while the TAG offered guidance on policy opportunities Additionally, two public outreach meetings were held at New York University's Puck Building on June 16, 2009, and December 8, 2009, to discuss key issues surrounding OHD These meetings aimed to identify specific concerns and assess the implications of current truck traffic and potential changes due to the project.

3.2.2.1 The role of financial incentives

A study by Holguin-Veras highlights that the effectiveness of delivery programs varies based on who bears the costs or receives incentives Implementing a pricing strategy, such as higher tolls for daytime deliveries, is unlikely to significantly encourage a shift to off-peak or nighttime deliveries, as receivers typically dictate delivery time windows While off-hour deliveries can offer carriers advantages like reduced travel time, fewer trucks, lower fuel and traffic violation costs, and overall logistics savings, receivers gain little from this shift Without substantial benefits for receivers, there is little motivation for them to voluntarily opt for off-hour deliveries.

Shifting to off-hours delivery may not be financially beneficial for businesses, as it requires additional employees to manage deliveries However, offering financial incentives to recipients could motivate them to engage in off-hours delivery.

To ensure participation in OHD, it is crucial that the benefits outweigh the costs for carriers, shippers, and receivers The project team focused on identifying effective incentives, examining both broad-based and targeted financial incentives for receivers, as well as a voluntary green certificate program to encourage involvement.

The project team identified that offering financial incentives of $5,000 to $10,000 to receivers could significantly encourage off-hour deliveries, potentially shifting 3-7% of deliveries during peak congestion times With an estimated value of $25 per hour for reduced congestion, the benefits of this policy would outweigh its costs However, the complexity of coordinating among multiple receivers and carriers may pose challenges, potentially increasing the overall implementation costs of the program.

Incentives aimed at Large Traffic Generators (LTGs) are likely to attract more interest from practitioners and decision-makers due to the complexities of receiver-carrier coordination within broad-based programs By focusing on LTGs, which include clusters of businesses like malls or large buildings with distinct ZIP codes, significant shifts in truck traffic to off-peak hours can be achieved These facilities typically produce high volumes of truck traffic and manage both incoming and outgoing deliveries through a central delivery station, resulting in a substantial payoff for off-hour traffic initiatives.

To enhance off-hour delivery in limited time groups (LTGs), it is essential to implement tailored information campaigns for building owners An IAG member highlighted that, in their building with 103 retail tenants, off-hour deliveries (OHD) are permitted only when tenants have designated staff to receive them Therefore, allowing unassisted deliveries could significantly encourage the adoption of OHD practices.

The "Green Certificate" model serves as a non-monetary incentive that enhances a company's image, encouraging businesses to act as responsible corporate citizens By voluntarily adopting sustainable practices, companies can boost their reputation, ultimately leading to increased customer loyalty and sales Public agencies could establish a Green Certificate program to acknowledge these efforts, offering varying levels of certification based on participation This public recognition not only motivates companies to improve but also allows customers to advocate for sustainability among non-compliant businesses The implementation of such a program in a major city like New York could have significant influence and promote widespread change.

A purely voluntary program without incentives currently lacks measurable benefits, as those willing to participate would already be engaged under the existing conditions Introducing a program with recognition, such as a "Green Apple" award, could potentially influence participation, although the exact impact remains uncertain Overall, while the benefits of such a program are likely to be positive, the associated costs would be minimal.

3.2.3 Foreseeable barriers that may incur costs and suggested solutions

3.2.3.1 Extra employees for off-hours: roving crews and/or tax credits

The behavioral model of this study suggests potential extra costs for receivers, but the appropriateness of these costs remains uncertain until a program is implemented A significant expense associated with implementing Overnight Delivery (OHD) is labor Retail stores that do not operate 24 hours would require hiring an additional employee specifically for handling deliveries, which poses a challenge given the limited delivery window of just one to two hours As noted by an IAG member, labor costs can be substantial, with some companies reporting nearly a 10% cost differential for employees working overnight shifts.

Interviews with IAG members revealed two potential solutions for reducing labor costs related to Off-Hours Deliveries (OHD) The first option involves the implementation of "roving crews" that would serve multiple stores within specific commercial areas, such as malls or buildings, allowing for shared labor expenses The second option is the introduction of a tax credit aimed at offsetting labor costs This idea has gained some traction, particularly with U.S Representative Anthony Weiner from New York's 9th District, who has proposed city tax credits to support businesses transitioning to off-peak delivery times He suggests that these businesses could receive city tax credits, which would be matched by federal funds, to help cover staffing expenses (Weiner, 2010).

Retailers face a significant challenge in staffing for off-hours to receive goods, which incurs extra expenses related to wages, heating, lighting, and insurance One effective solution to this labor dilemma is the implementation of unassisted delivery facilities, such as central drop box locations.

However, the implementation of a central drop box strategy would also entail some additional costs

To facilitate the installation of a central drop box, modifications to building codes may be necessary Furthermore, public sector subsidies could be essential for the construction of these facilities Future studies by the project team should assess the acceptance of unassisted delivery options and include performance simulations that consider associated building costs.

Noise pollution is a significant issue driving neighborhood resistance to Out-of-Hours Deliveries (OHD) in Europe, particularly highlighted in Athens, Greece, where nighttime unloading operations create substantial disturbances (Yannis et al., 2006) This concern is echoed by IAG members, emphasizing the complexities of managing noise in mixed-use areas where businesses and residences coexist Solutions may involve amending local ordinances to regulate noise levels during specific hours (Ogden, 1992) or implementing noise reduction strategies, such as using plastic roll cages instead of metal ones to minimize noise during deliveries (Whitegift Centre, 2007) However, as noted by Browne et al., these noise abatement measures can lead to increased costs for both receivers and carriers (Browne et al., 2006).

Institutional Arrangements Suggested

3.3.1 Types of potential institutional arrangements

The implementation of OHD (On-Demand Freight Delivery) offers numerous advantages, including lower logistics costs, reduced travel time, decreased traffic congestion, and improved air quality However, its success hinges on establishing new institutional arrangements that foster collaboration between public and private sectors A report by the National Cooperative Freight Research Program defines such arrangements as foundational structures that facilitate the advancement of freight mobility through coordinated infrastructure, operations, services, and regulations.

39 into three types depending on specific objectives of each arrangement First, Type I organizations focus on the advocacy function by concentrating on education and consensus building (Cambridge Systematics,

The 2009 study has significantly enhanced collaboration between public and private sector stakeholders by establishing the IAG and TAG, as well as conducting public outreach meetings to promote the concept of OHD Type II organizations are responsible for evaluating, prioritizing, and funding freight projects within specific regions or types, such as city or state transportation agencies Additionally, Type III organizations are created to execute specific projects, handling aspects like financing, environmental clearances, and contractual negotiations.

The evolution of organizations from Type I to Type III reflects a linear development process, where Type III organizations encompass the characteristics of the previous types Typically, projects initiate with needs identification and stakeholder consensus (Type I), followed by alternative analysis for prioritization within public agencies (Type II), and culminate in the establishment of a large institution for implementing extensive projects through multi-jurisdictional collaboration and private sector partnerships (Type III) The Alameda Corridor project exemplifies this progression in freight transportation planning, showcasing the collaboration among various agencies and private stakeholders, including the Southern California Association of Governments, city and state governments, and private entities like truckers and rail operators Such public-private partnerships could similarly facilitate the successful implementation of an OHD program in the New York metropolitan area.

The successful establishment of a new institutional arrangement requires four essential conditions Firstly, it necessitates champions who possess vision, commitment, drive, and initiative, often supported by political will to educate stakeholders and implement an OHD scheme This process may be initiated through top-down legislative mandates or bottom-up approaches by technical staff Given the significant interest from public agencies in the New York metropolitan area, the next step is to create a formal policy that guides OHD implementation Secondly, identifying key private sector stakeholders is crucial, along with developing incentive programs to attract them Research indicates that many private sector participants believe OHD cannot be effectively implemented without such incentives.

2006) Third, the needs assessment workshop with stakeholders is useful to define the purpose of a new institutional arrangement (Cambridge Systematics, 2009) While the two public outreach meetings for the

A recent study has highlighted ongoing challenges in implementing Occupational Health and Safety (OHD) policies, indicating a need for a more comprehensive roundtable to identify stakeholders and establish a new institutional framework Following preliminary consensus among the exploratory group on the necessity and objectives of this institutional arrangement, it is crucial to develop an action plan outlining both short-term and long-term activities (Cambridge Systematics, 2009).

The project team engaged with key community leaders and firms to assess reactions to the concept of OHD While there is considerable support for OHD, concerns about excessive government intervention in private operations persist The costs associated with transitioning to OHD differ significantly among firms, locations, and other factors, making a uniform incentive level potentially problematic, leading to either overpayment or underpayment that could waste resources or undermine the program's effectiveness Additionally, a cost-specific program for each firm would be challenging to manage and susceptible to misuse Although further research is needed, initial interactions with local stakeholders have provided valuable insights.

Robust public sector support is crucial for the success of any voluntary program, particularly in urban settings Recognition and endorsement from city government play a vital role, especially given the multiple agencies involved in supporting OHD shifts Effective coordination among local government entities—such as traffic, environmental, planning, transit, and sanitation—requires a committed leader backed by the Mayor and local officials to ensure the program's success.

Effective leadership in the public sector must be complemented by strong involvement from the private sector to ensure the success of either the voluntary program or the targeted LTG program Ongoing discussions with various private sector stakeholders, including freight companies, goods receivers, and landlords, highlight the need for increased engagement in the implementation process Encouragingly, major firms have expressed a positive outlook, with some ready to transition to OHD, and a prominent real estate association showing interest in the program on behalf of landowners.

In the current challenging financial climate, ongoing operating subsidies for numerous firms are unlikely to be sustainable due to depleted government resources following the recent recession Public authorities are compelled to make significant service reductions and cuts to capital programs, with tolls and fare increases already allocated to support essential operations Consequently, any additional financial commitments from the public sector should be limited to one-time support rather than ongoing funding.

Building owners and tenants should prioritize physical improvements to facilitate unstaffed receiving locations and enhance security Before planning any large-scale subsidies, it's crucial to consider a robust voluntary program that can effectively demonstrate the potential for meaningful change Successful implementation of this program could generate demand for larger initiatives and assist the public sector in accurately assessing the financial implications of transitioning to on-demand delivery for goods receivers.

References

Browne, M., J Allen, S Anderson and A G Woodburn (2006) Night-time Delivery Restrictions: A review Recent Advances in City Logistics: Proceedings of the 4th International Conference on City Logistics, Amsterdam, Elsevier Ltd

Cambridge Systematics (2009) Institutional Arrangements for Freight Transportation Systems Transportation Research Board http://onlinepubs.trb.org/onlinepubs/ncfrp/ncfrp_rpt_002.pdf

Holguín-Veras, J (2006) Potential for Off-Peak Freight Deliveries to Congested Urban Areas NYSDOT http://www.rpi.edu/~holguj2/OPD/OPD_FINAL_REPORT_12-18-06.pdf

Holguín-Veras, J (2008) "Necessary conditions for off-hour deliveries and the effectiveness of urban freight road pricing and alternative financial policies in competitive markets." Transportation Research Part A: Policy and Practice 42(2): 392-413

Holguin-Veras, J., K Ozbay and A de Cerreno (2005) Evaluation Study of the Port Authority of New York and New Jersey's Time of Day Pricing Initiative NJODOT

Huschebeck, M (2004) BEST Urban Freight Solutions: Deliverable D1.4 - Recommendations for Further Activities Best Urban Freight Solutions (BESTUFS)

NICHES (2008) "Innovative Approaches in City Logistics: Inner-City Night Delivery." Retrieved March

26, 2010, from http://www.niches- transport.org/fileadmin/archive/Deliverables/D4.db_5.8_b_PolicyNotes/14683_pn7_night_delivery_o k_low.pdf

NYMTC (2007) The Basics of Freight Transportation in the New York Region New York Metropolitan Transportation Council

Ogden, K W (1992) Urban Goods Movement: A Guide to Policy and Planning Brookfield, VT, Ashgate Publishing Company

In 2010, Roberts emphasized the importance of effective institutions for transportation operations, providing recommendations for reauthorization to enhance efficiency (Roberts, D C.) Similarly, Weiner proposed strategies for reducing congestion without imposing regressive taxes, highlighting innovative solutions for urban transportation challenges (Weiner, A.) Both articles underscore the need for strategic planning and policy reform to improve transportation systems.

The Whitegift Centre (2007) highlights the importance of nighttime deliveries in South London as a sustainable solution to urban logistics challenges This approach aims to minimize traffic congestion and enhance delivery efficiency Additionally, research by Yannis, Golias, and Antoniou (2006) examines the impacts of urban delivery restrictions on traffic movements, emphasizing the need for effective transportation planning to balance urban development and delivery operations Together, these studies underscore the significance of innovative delivery strategies in promoting sustainable urban environments.

MARKET ANALYSES

Identification and Quantification of Potential Target Markets

Implementing off-hour delivery programs requires careful selection of target industry segments to ensure effectiveness Previous research highlights that a minimum scale of operation is essential for carriers to benefit from off-hour deliveries (OHD) By concentrating on specific industries, carriers can achieve the necessary scale to make OHD viable and advantageous.

This section outlines the methodology for identifying optimal market segments for participation in OHD in Manhattan Previous studies have primarily utilized commodities as indicators for market segmentation; however, this approach has limitations, particularly in linking model outcomes to specific business types due to the exclusive focus on commodity types This disconnect makes it challenging to align behavioral models with economic databases classified by business lines, such as the Standard Industrial Classification (SIC) system, which is crucial for analyses like trip generation Additionally, the complexity arises from businesses receiving multiple commodities from various carriers, complicating the determination of commodity distribution In contrast, employing SIC codes facilitates the connection between survey data and information on business establishments in the New York City metropolitan area.

Identifying the optimal market segments for off-hour delivery (OHD) is essential for reducing congestion, as the relationship between carriers and receivers significantly influences delivery times Carriers are motivated to increase off-hour deliveries due to potential transportation cost savings, but the success of OHD hinges on encouraging receivers to request these deliveries Since receivers control the delivery schedule, understanding which ones are open to shifting to OHD is vital While carriers benefit from lower costs, receivers often face increased expenses due to limited operating hours, making them hesitant to adapt Therefore, pinpointing the most receptive receivers is crucial for developing effective policies that promote OHD.

The approach used utilizes the data previously obtained from a New York State Department of Transportation funded project, which includes descriptive characteristics of the businesses surveyed, such

The study examines the number of employees and deliveries, alongside their expressed preferences regarding the feasibility and willingness to transition to On-Demand Delivery (OHD) under different policy scenarios.

4.1.1 Behavioral modeling of off-hour delivery initiatives

To identify industries more inclined to adopt off-hours deliveries (OHD), behavioral modeling was conducted using discrete choice models based on scenarios outlined by Holguin-Veras et al (2007; 2008) These scenarios aimed to evaluate which policies would most effectively encourage participation from both carriers and receivers in OHD initiatives.

This section presents modeling results divided into two sub-sections: one for receivers and another for carriers For receivers, the analysis includes scenarios such as tax deductions for opting for off-hour deliveries (OHD) and reduced shipping costs during off-peak times In contrast, the carrier scenarios focus on customer demand for OHD and various incentives, including designated street-side parking during off-hours, pre-approved security clearances for trucks at tunnels and bridges, toll savings, financial rewards per mile for off-hour travel, and permits for OHD.

The most effective models integrate fundamental company characteristics, interaction terms, and behavioral variables, including facility type, employee count, and primary business sector A positive coefficient signifies a beneficial relationship between the variable and the utility of implementing OHD, whereas a negative coefficient indicates a detrimental effect Additionally, a bootstrap process was employed to address the repeated measurement issue inherent in the data.

Two policy scenarios aimed at receivers were evaluated: (1) implementing a tax deduction and (2) offering shipping cost discounts for those accepting off-hour deliveries (OHD) These receiver-focused policies provide financial incentives to offset the additional expenses incurred due to off-hour operations.

A recent survey explored the willingness of receivers to accept a percentage of their deliveries during off-hours in exchange for tax deductions of $3,000, $6,000, and $9,000 annually The findings revealed a significant positive correlation between the tax deduction amount and the likelihood of accepting off-hour deliveries (OHD) Furthermore, the analysis showed that the type of business also influences this likelihood, with certain business types demonstrating a higher propensity to accept OHD when offered a tax incentive.

44 involved in building materials, home furnishings, retail baked goods, and eating establishments while the likelihood decreases for receivers of jewelry

In a survey examining off-hour delivery (OHD) acceptance, participants were asked about their willingness to receive deliveries during off-hours in exchange for reduced shipping costs of 20% and 40% The results indicated a significant positive correlation, showing that as the shipping cost savings increased, so did the likelihood of receivers accepting OHD Additionally, receivers from sectors such as home furnishings, professional and commercial equipment and supplies, piece goods, and liquor stores demonstrated a higher propensity for accepting off-hour deliveries.

This section presents the outcomes of various policy scenarios aimed at carriers, focusing on the percentage of customers requesting Off-Hours Deliveries (OHD) Key incentives include designated street-side parking, pre-approved security clearances for trucks to avoid inspections at tunnels and bridges, toll savings, financial rewards for off-peak travel, and permits for OHD Additionally, the potential use of a neutral company for the final leg of deliveries to Manhattan is examined, although it is not classified as a policy scenario.

The analysis of the first carrier scenario highlights the influence of three key incentives on the likelihood of participation in off-hour deliveries (OHD): a specified percentage of customers requesting OHD (25%, 50%, or 75%), designated street-side parking for OHD, and pre-approved security clearances for trucks to bypass inspections The findings indicate that all three incentives are significant, with the percentage of customers requesting OHD being particularly impactful Designated street-side parking during off-hours notably increases the likelihood of OHD participation, especially for carriers facing average monthly fines under $400, as parking issues are less critical for them Additionally, carriers transporting wood/lumber and medical supplies are more likely to engage in OHD, while long-haul trucking firms show a greater inclination to participate compared to local trucking firms.

The second carrier scenario examined the potential for carriers to increase the number of Off-Hours Deliveries (OHD) based on customer demand and toll savings for using bridges and tunnels during off-peak times Toll savings were established at $3, $4, or $7 per axle The model indicated that the percentage of customers requesting OHD significantly influenced the likelihood of carrier participation, demonstrating a positive correlation between customer demand and carrier engagement in the OHD program.

In the analysis, it was found that carriers receiving toll discounts and operating in the grocery, furniture, or non-local trucking sectors are more likely to engage in OHD Additionally, various business line variables, such as Shipper, Warehouse, and Mover, demonstrated significant positive correlations with the likelihood of participating in OHD.

The third carrier scenario assessed the potential for carriers to increase off-hour deliveries (OHD) to Manhattan, contingent on customer demand and financial incentives based on mileage traveled during off-peak times The study offered financial rewards of either 5 cents or 10 cents per mile Findings revealed that carriers involved in food transport are more likely to engage in this scenario, while non-local carriers show a greater willingness to make OHD due to longer travel distances, allowing them to maximize their benefits from off-hour deliveries.

Behavioral Analyses: Behavioral Micro-Simulation (BMS)

Section 4.1 discussed the results of various models to determine the impact of various policies on the behavior of receivers and carriers It also provided an estimate of the number of daily freight deliveries made to Manhattan This section utilizes the results of the behavior models to create the Behavioral Micro-Simulation (BMS) model which determines the shift in deliveries to the off-hour in various industry segments for various incentive levels

This model aims to enhance participation in OHD by utilizing a BMS to analyze carrier-receiver interactions, ultimately assessing its effectiveness.

The article discusses various policy incentives, including time-of-day pricing and financial rewards for receivers who accept On-Demand Home Delivery (OHD) A key aspect of the Behavioral Management System (BMS) involves simulating the behaviors of both receivers and carriers The simulation for receivers employs a discrete choice model that incorporates tax deduction incentives to predict their acceptance or rejection of OHD Consequently, carriers analyze the receivers' decisions and the subsequent effects on delivery costs to determine their own strategies regarding OHD implementation.

Tax deductions for receivers who extend their operational hours can significantly boost off-hour delivery (OHD) participation, particularly in industries such as Food, Non-Alcoholic Beverages, Alcoholic Beverages, Wood/Lumber, Paper, Chemicals, Plastic, and Medical Supplies Analysis using the BMS indicates that carriers located near urban customers are more inclined to engage in OHD, as they can quickly reap the benefits Furthermore, the study suggests that stricter enforcement of parking fines for double parking during regular hours could encourage more carriers to participate in OHD, and that financial incentives, including tax deductions for receivers and rewards for carriers, have a greater impact on OHD than toll surcharges.

4.2.1 Scope of the behavioral analyses

Previous research on off-hour delivery participation has primarily utilized discrete choice models to estimate market shares; however, this approach has notable limitations These models fail to adequately consider geographic factors that influence participation decisions, such as the distances of delivery routes in relation to receiver locations.

Delivery costs vary throughout the day due to the geographic locations of both receivers and carriers, influencing logistics operations and routing This variability in costs is crucial, as it significantly impacts the decision-making process for scheduling deliveries in the On-Demand Delivery (OHD) model.

To effectively assess the impact of off-hour delivery (OHD) operations, it is essential to incorporate various real-world factors into the analysis Key considerations include budget and working constraints faced by both receivers and carriers, the selection of delivery routes, and productivity levels, which encompass travel time and speed estimates during both regular and off-peak hours Additionally, the influence of government regulations and policies that may restrict off-hour delivery activities must be evaluated A comprehensive understanding of OHD effectiveness requires a simultaneous examination of these factors alongside potential financial incentives, such as tax deductions and time-of-day tolls.

49 following sections describe the BMS developed by the authors to address these limitations, and to discuss preliminary findings achieved by applying it to a set of hypothetical test cases

The Behavioral Micro-Simulation (BMS) framework, illustrated in Figure 4, outlines a sequential decision-making process that impacts participation in Off-Hours Deliveries (OHD) (Holguín-Veras et al., 2008) Central to this simulation are two key modules: the receiver behavioral simulation and the carrier behavioral simulation The decisions to engage in OHD are primarily shaped by policy incentives provided to both receivers and carriers, represented as Πr and Πc Additionally, carriers' choices regarding OHD are directly affected by receivers' acceptance of these deliveries and the associated delivery costs Ultimately, the BMS aims to analyze how economic incentives for receivers and carriers can enhance carrier participation in OHD, thereby reducing truck traffic in urban areas during both peak and off-peak hours.

Figure 4: Behavioral Micro-Simulation (BMS) Framework

Receivers benefit from a tax deduction for accepting deliveries during off-hours, while carriers face time-of-day tolls with surcharges for regular-hour deliveries These strategies were chosen for micro-simulation due to their proven effectiveness in promoting off-hour deliveries (OHD) (Holguín-Veras et al., 2007; 2008) The study aims to analyze how carrier participation in OHD shifts when receivers are incentivized to adjust their receiving schedules.

The carrier-receivers selection process is designed to create a synthetic population of carriers to analyze how delivery scheduling of commodities is affected by incentives This process considers various industry segments, including plastics, jewelry, chemicals, and food, among others (Holguín-Veras et al., 2008) It begins by randomly selecting an industry segment, followed by choosing a carrier from that segment's population Subsequently, the number of stops made by the selected carrier determines how many receivers are chosen from the corresponding industry segment This group of receivers will be used for behavioral simulations, enabling the carrier to estimate delivery costs during both regular and off-hours, ultimately informing decisions regarding participation in off-hour activities.

The micro-simulation of receiver behavior is based on a behavioral model that incorporates the impact of tax deductions, as detailed in Table 6 This model was developed using data from a project funded by the New York State Department of Transportation, which explored the potential for off-peak freight deliveries in congested urban areas In 2005, interviews with 200 receivers in Manhattan provided insights into their receiving and shipping patterns, operational flexibility, willingness to adopt off-hour deliveries (OHD) with economic incentives, and company characteristics The model considers the tax deduction variable, reasons for rejecting OHD, and interaction terms related to various commodity types, including wood, alcohol, paper, medical supplies, food, printed material, and metal.

Table 6: Binary Logit Model for Receiver Tax Deduction Scenario (Holguín-Veras et al., 2008)

Utility of off-peak deliveries: C1CHOICE

A tax deduction for an employee assigned to OHD TDEDUCT 8.392E-05 1.410

Reasons for not receiving OHD

No access to building/freight entrance after hours REASON1 -1.234 -1.571

Additional costs to the business if accepting more OHD COST -0.888 -3.232

Interferes with normal business REASON2 -0.591 -1.208

Tax deduction for receivers of Wood/lumber TDCOM8 6.968E-04 2.219

Tax deduction for receivers of Alcohol TDCOM4 4.356E-04 2.209

Tax deduction for receivers of Paper TDCOM9 2.627E-04 2.988

Tax deduction for receivers of Medical supplies TDCOM22 2.598E-04 3.188

Tax deduction for receivers of Food TDCOM2 1.875E-04 3.973

Tax deduction for receivers of Printed Material TDCOM21 1.652E-04 1.802

Tax deduction for receivers of Metal TDCOM13 1.415E-04 1.410

Number of employees in a branch facility BRANEMP 9.867E-03 1.612

Utility of no off-peak deliveries:

The BMS simulates the response of receivers to a given incentive by computing the utility of accepting the incentive, U(OHD), using discrete choice models This calculation determines the probability of acceptance, P(OHD), for each receiver If P(OHD) exceeds a randomly generated number between 0 and 1, the receiver accepts the incentive This process is repeated for all selected receivers across various incentive levels, providing a comprehensive assessment of the incentive's effectiveness.

The third component of the simulation involves carrier behavioral simulation, where carriers make decisions based on the incentives accepted by receivers for OHD These decisions are influenced by the profits generated, starting with the selection of receivers tailored to each carrier's industry segment The number of receivers is determined by the carrier's number of stops, and their geographic locations establish the base case conditions for calculating delivery costs, as illustrated in Figure 5.

To estimate what would happen as a response to the incentives, the receivers‘ decisions are simulated

Incentives like tax deductions can lead some recipients to opt for off-hour deliveries, while others may prefer regular delivery times This shift results in the division of the original network into two distinct sub-networks, reflecting a mixed case scenario.

The analysis involves 52 deliveries during regular and off-hours, as illustrated in Figures 6 and 7 For instance, when a carrier makes five stops, five receivers are randomly chosen If two of these receivers accept off-hour deliveries (OHD) while three do not, it is essential to estimate optimal routes for both off-hour and regular-hour deliveries Subsequently, the transportation costs for regular-hour deliveries must be calculated.

Behavioral Analyses: Approximation model

This section aims to create an approximation model that estimates the joint carrier-receiver response to OHD policies, simplifying the process by eliminating the need for costly data-driven calibration methods By providing accessible models, transportation agencies and metropolitan planning organizations can more effectively analyze and design OHD programs and policies.

This paper largely adheres to the notation established in prior research by Holguín-Veras (2008), with subscripts i and j denoting receiver i and carrier j, respectively Additionally, the superscripts BC, R, and O represent base case, regular, and off-hour operations.

G j , G M j = Gross revenues (base case, mixed operation) to carrier j

D ( ) = Incremental gross revenues to carrier j associated to policy  C

C j = Total cost of carrier j‘s base case operations (no off-hour deliveries)

C   = Total cost of carrier j‘s mixed operations (regular plus off-hour deliveries)

C j = Total cost of carrier j associated with regular deliveries in a mixed operation

C j = Total cost of carrier j associated with off-hour deliveries in a mixed operation

D ( ) = Incremental total costs to carrier j in response to policy  C

D ( ) = Incremental gross revenues to receiver i associated to policy  R

D ( ) = Incremental total costs to receiver i associated with switch to off-hours in response to policy  R j

D = Incremental fixed costs to carrier j j

D = Incremental distance costs to carrier j j

D = Incremental time costs to carrier j j

D = Incremental toll costs to carrier j

C FC , C FC R , C FC O = Cost of trip to first customer (base case, regular, and off-hour operations)

C HB , C HB R , C HB O = Cost of returning to home base (base case, regular, and off-hour operations)

BC c D , c D R , c D O = Unit cost per distance traveled (base case, regular, and off-hour operations)

BC c T , c T R , c T O = Unit cost per time traveled (base case, regular, and off-hour operations)

D BC , D R , D O = Tour distance (base case, regular, and off-hour operations)

S R = Toll surcharge to trucks traveling during regular hours as part of the cordon scheme

 D ,  O D = Distance based unit toll for distance traveled in tolled area (regular, and off-hours)

 T ,  T O = Time based unit toll for time spent in tolled area (regular, and off-hours)

 i = Length of time during which off-hour deliveries are accepted by receiver i

min = Minimum amount of time required for off-hour deliveries

A = service area, i.e., area of the minimum size rectangle that envelopes all customers max max y x L

A = Size of the actual service area

L ox = X dimension of the rectangular service area

L oy = Y dimension of the rectangular service area oy ox o L L

N   = Total number of customers for base case conditions

N R , N O = Total number of customers during regular and off-hours (mixed operation) u R , u O = Average travel speeds (regular and off-hours)

 = Ratio of average travel speeds

 = Ratio of unit time costs

 = Original set of receivers during base case conditions, served by carrier j

 j = Set of receivers, served by carrier j, that prefers regular-hour deliveries

 j = Set of receivers, served by carrier j, that decides to accept off-hour deliveries

 O = set of carriers that do off-hour deliveries

F = Financial incentive provided to receivers for committing to accept off-hour deliveries

P = Probability that a receiver would commit to off-hour deliveries

4.3.2 Carrier-receiver interactions, necessary conditions, and impacts of pricing

The formulation of the approximation model relies on key analytical developments that support its foundational assumptions, particularly the necessary conditions for carriers and receivers to switch to off-hours (Holguín-Veras, 2008) and research on the effects of cordon time-of-day and time-distance pricing (Holguín-Veras, 2009) This paper emphasizes that the interactions between carriers and receivers significantly influence their joint responses to pricing strategies Carriers generally prefer off-hour deliveries (OHD) for increased productivity and reduced costs, while most receivers favor regular-hour deliveries, which align with their existing staffing and avoid extra expenses This dynamic can be likened to the Battle of the Sexes game (Rasmusen, 2001), characterized by two Nash equilibria, where the outcome is determined by the more powerful player Given that most deliveries occur during regular hours (Holguín-Veras et al., 2007), it is evident that receivers hold the dominant position in this interaction.

The consideration of carriers and receivers as distinct economic agents enhances the understanding of their interactions regarding delivery times and pricing decisions (Holguín-Veras, 2009) This model effectively clarifies the behavioral responses to pricing changes, as evidenced by carriers' reactions following the Port Authority of New York and New Jersey's implementation of time-of-day pricing Carriers primarily sought productivity increases to manage toll hikes, could only pass on toll costs to customers in 9% of instances, and cited "customer requirements" as the reason for their inability to adjust behaviors in 70% of cases (Holguín-Veras et al., 2006c) Such behaviors are comprehensible only when considering carrier-receiver interactions within a competitive market framework.

For carrier-receiver interactions to transition to off-hours, both parties must experience improved conditions This relationship can be mathematically expressed in relation to policies targeting carriers and receivers, as outlined by Holguín-Veras (2008).

In the context of policy  R, the incremental gross revenues (DG i) and costs (DC i) for receiver i are analyzed alongside the incremental gross revenues (DG j) and costs (DC j) for carrier j under policy  C Additionally, the delivery time for receiver i is denoted as  i O, highlighting the impact of off-hour shifts on both receivers and carriers.

The equations form the foundation for creating cost functions that reflect the expenses carriers incur when delivering to a group of N receivers, categorized into regular and off-hours These cost functions are essential for estimating delivery rates and determining the extent to which carriers can transfer toll costs to the receivers.

In a study by Holguín-Veras (2009), the interplay between carriers and receivers regarding pricing and carrier-receiver policies was examined, revealing that financial incentives could prompt some receivers to shift to off-hours, resulting in a mixed operational model for carriers The research categorizes profitability into three scenarios: quasi-best, expected value, and worst case Optimal tour distances were estimated using a model for the Probabilistic Traveling Salesman Problem, and the analytical cost functions incorporated fixed costs related to travel to and from the home base, as well as time, distance, and toll costs under both cordon time-of-day and time-distance pricing The findings detail incremental costs to the carrier, highlighting negative costs as potential savings, with specific subscripts denoting fixed, distance, time, and toll costs associated with different pricing strategies Equations (4) through (11) illustrate the cost functions for cordon time-of-day pricing.

4.3.2.1 Summary of results for cordon time-of-day pricing

All cases (quasi-best, expected value, worst case):

The incremental fixed costs remain constant across different scenarios, while the incremental costs related to distance, time, and tolls vary significantly Notably, there is often a discontinuity in costs when all receivers transition to off-peak hours.

The analysis indicates that a fixed cost arises from additional off-hour trips, except when all receivers are available during these hours, which eliminates the fixed costs This finding highlights that as the distance between the carrier and the delivery area increases, the incremental fixed costs also rise, making it increasingly challenging for mixed operations to achieve profitability.

Equation (5) reveals that the toll surcharge incentivizes carriers only when all receivers operate during off-hours, as this allows for a single tour that avoids tolls during regular hours In other scenarios, carriers must travel in both regular and off-hours, incurring tolls regardless, resulting in no incremental cost compared to the base case Consequently, this raises concerns about the effectiveness of cordon time-of-day pricing for managing freight demand.

The analysis reveals that the incremental costs in distance and time vary based on the scenario examined In the optimal scenario, even a slight shift of receivers to off-hours can yield savings in both time and distance, with the incremental distance cost potentially reaching zero Additionally, the incremental time cost remains negative when the ratio of wage increase to speed difference between off-hours and regular hours is less than one In scenarios with expected values, the problem's quadratic nature initially leads to rising costs, which eventually decrease, resulting in savings when a significant number of receivers operate during off-hours Conversely, in the worst-case scenario, increases in both distance and time costs are prevalent.

The analytical derivations for time-distance pricing reveal that the incremental fixed, distance, and time costs are identical for cordon time-of-day pricing, as demonstrated in equations (4) to (11) However, the primary distinction lies in the incremental toll costs, which are detailed below.

4.3.2.2 Summary of results for time-distance pricing

The results clearly show that time-distance pricing incentivizes carriers to operate during off-hours, as the unit distance tolls contribute to incremental toll costs Additionally, implementing a robust pricing policy, where the unit tolls for regular hours exceed those for off-hours, can result in cost savings for carriers, irrespective of the number of receivers during off-peak times Notably, a greater disparity between the unit tolls for regular and off-hours enhances these benefits.

The findings indicate that as the incentive for the carrier increases, there is a significant difference compared to cordon time-of-day pricing, where the toll surcharge is only activated when all receivers transition to off-peak hours.

4.3.3 Minimum number of receivers for a profitable mixed operation

Assessment of Impacts

After analyzing the freight delivery estimates for Manhattan and the subsequent changes in OHD's market share at different incentive levels set by the BMS, the findings were utilized to identify target markets for policies aimed at promoting OHD.

References

Beardwood, J., J H Halton and J K Hammersley (1959) "The Shortest Path Through Many Points." Proceedings of the Cambridge Philosophical Society 55: 299-328

Churchill, J D C (1970) Operation "MoonDrop" : An Experiment in out of Hours Goods Delivery Proceedings of the 3rd Technology Assessment Review, Paris, France

Holguín-Veras, J (2006) Potential for Off-Peak Freight Deliveries to Congested Urban Areas NYSDOT http://www.rpi.edu/~holguj2/OPD/OPD_FINAL_REPORT_12-18-06.pdf

Holguín-Veras, J (2008) "Necessary conditions for off-hour deliveries and the effectiveness of urban freight road pricing and alternative financial policies in competitive markets." Transportation Research Part A: Policy and Practice 42(2): 392-413

Holguín-Veras, J (2009) Urban Delivery Industry Response to Cordon Pricing, Time-Distance Pricing, and Carrier-Receiver Policies International Transportation Economics Conference Minneapolis

Holguín-Veras, J., M A Silas and J Polimeni (2006b) "On the Overall Performance of Comprehensive Policies to Manage Truck Traffic in Congested Urban Areas." First International Conference on Funding Transportation Infrastructure

In their 2007 study published in "Networks and Spatial Economics," Holguín-Veras et al investigate the effectiveness of joint receiver-carrier policies aimed at increasing truck traffic during off-peak hours The research focuses on the behaviors of receivers and highlights how strategic policy implementation can optimize logistics and improve traffic flow The findings suggest that incentivizing off-peak deliveries can lead to significant benefits for both carriers and receivers, ultimately enhancing overall transportation efficiency.

Holguín-Veras et al (2008) explore the effectiveness of joint receiver-carrier policies aimed at increasing truck traffic during off-peak hours, highlighting the behaviors of carriers in response to these strategies In a related study, Holguín-Veras et al (2006) analyze how time-of-day pricing influences freight carrier behavior in congested urban settings, discussing the broader implications for road pricing policies Both studies contribute valuable insights into optimizing freight movement and reducing congestion in urban transportation networks.

Kjaersgaard, S and H Enslev Jensen (2003) Sustainable City Logistic Solutions City Logistics III, Madeira, Portugal

Ogden, K W (1992) Urban Goods Movement: A Guide to Policy and Planning Brookfield, VT, Ashgate Publishing Company

Port Authority of New York and New Jersey (2009) "PANYNJ 2008 Toll Schedule." Retrieved March

23, 2009, from http://www.panynj.gov/COMMUTINGTRAVEL/tunnels/html/tolls.html

Rasmusen, E (2001) Games and Information: An Introduction to Game Theory Malden, Massachusetts, Blackwell Publishers

Silas, M and J Holguín-Veras (2009) "A Behavioral Micro-Simulation for the Design of Off-Hour Delivery Policies." Transportation Research Record 2097: 43-50

PILOT TEST PREPARATIONS

Process Followed to Assemble the Companies for the Pilot Test

The team's pilot test faced significant delays due to the financial industry meltdown on Wall Street, which severely affected New York City's economy and made it difficult to recruit participants The collapse highlighted the critical role of the financial sector in generating 20% of New York State's tax revenues and its impact on local restaurants that relied on financial sector customers As a result, the downturn not only harmed the state's economy but also particularly affected the restaurant sector, which was the focus of the pilot test Consequently, the team had to frequently adjust its approach to successfully organize the test amidst this challenging economic landscape.

Phase I: Recruitment through the Industry Advisory Group (February-April 2008)

In February 2008, efforts intensified to recruit participants for the pilot test, following a preparatory phase from June to December 2007 that involved establishing the Industry Advisory Group and Agency Advisory Group, collecting stakeholder input, assembling simulation networks, and reviewing relevant remote sensing technologies Companies were contacted for participation in the pilot test at the end of January 2008, providing them with adequate time to prepare for the holiday sales season and complete their post-holiday tasks.

The initial strategy involved utilizing the Industry Advisory Group (IAG) for participant recruitment in the pilot test However, the New York University team faced challenges in securing participants through outreach to the New Jersey and New York Motor Truck Associations This recruitment struggle occurred during an economic downturn, marked by the financial instability that followed the collapse of Bear Stearns, which significantly impacted New York City's economy.

Phase II: Recruitment via the New York State Restaurants Association (May-July 2008)

After recognizing stagnation in recruitment efforts via the IAG, a strategic shift was initiated The team collaborated with the New York State Restaurants Association (NYSRA) to solicit participation from its members Despite investing significant time and resources in crafting the participation request and conducting a survey among restaurants, there were no expressions of interest.

Phase III: Direct recruitment of restaurants (July-September 2008)

Facing challenges with recruitment through NYSRA, the team shifted strategies to directly reach out to restaurant owners They acquired a dataset with contact information for approximately 300 restaurants and sent out participation requests, followed by phone calls from team members Unfortunately, these efforts yielded no positive responses, coinciding with a significant financial crisis that resulted in the largest single-day drop in the Dow Jones.

Phase IV: No recruitment efforts (September-October 2008)

In the wake of Wall Street's financial meltdown, it became evident that economic instability was a key factor hindering the recruitment of participants for the pilot test Consequently, the decision was made to postpone recruitment efforts until the economy stabilized to a satisfactory level.

Phase V: Recruitment through “Project Champions” (November 2008-April 2009)

The significant influx of federal aid into the financial system has stabilized Wall Street, prompting the team to adopt an aggressive recruitment strategy This new approach targets "project champions," industry leaders who support off-hour deliveries (OHD) to encourage broader participation in the pilot test Coupled with a favorable economic climate, this strategy has successfully identified several promising candidates.

Union Square Hospitality Group: An influential restaurant group with twenty-five high end restaurants in New York City

B R Guest Restaurants: An important restaurant group with thirteen high end restaurants in

White Rose Food: A food distributor that delivers to more than 100 supermarkets in Midtown and Downtown Manhattan

Sysco: One of the largest distributors of food products in the United States It delivers food products to hundreds of food stores in New York City

Whole Foods Market: A large natural and organic grocery store chain with six stores in Manhattan

Foot Locker, a leading sportswear and footwear retail chain in the country, has partnered with New Deal Logistics, a company renowned for its commitment to environmental initiatives As the logistics provider for Foot Locker in Manhattan, New Deal Logistics sought collaboration on a new project, highlighting their dedication to sustainability within the retail sector.

The team ultimately chose to concentrate on Sysco, Whole Foods Market, and Foot Locker/New Deal Logistics, as these companies represent a well-rounded mix of industries, including restaurants (Sysco's clientele), grocery stores (Whole Foods Market), and retail (Foot Locker).

78 industrial partners of the project The next step was to proceed to plan the implementation of the pilot test, which is discussed next

Phase VI: Preparations for the Pilot Test (May-December 2009)

The project team successfully identified industrial partners for the pilot test and coordinated efforts based on whether participants were carriers or receivers For receivers like Whole Foods Market and Foot Locker, the team reached out to suppliers to encourage participation Notably, coordination with Foot Locker was minimal due to New Deal Logistics being the sole delivery provider In contrast, the team contacted 94 vendors serving Whole Foods Market in Manhattan, with 20 expressing interest in the pilot Key reasons for vendor reluctance included reliance on third-party delivery services, existing off-hour delivery schedules, and the challenge of managing split routes Ultimately, the team collaborated with the four participating Whole Foods Market stores and their vendors to adjust delivery times accordingly.

In midtown Manhattan, Sysco team members engaged with approximately 160 customers, resulting in 41 expressing interest in a pilot test, with 16 showing strong enthusiasm Among those not interested, common reasons included limited availability of staff during off-hours, building constraints that rendered OHD impractical, and the scale of their operations not warranting a change in their Sysco supply processes.

After establishing initial contacts, the company proceeded to finalize arrangements with its business partners—vendors for Whole Foods Market and food stores and restaurants for Sysco By mid-December 2009, they successfully completed this process, paving the way for the pilot test.

Phases of the Pilot Test

The pilot test involved four major industrial partners, each of whom shifted their entire distribution chains, particularly transportation and receiving operations, to off-hours for at least a month A total of 25 receivers, or 30 if including partial participants, and eight vendors took part in the test Due to the lack of interactions among the industrial partners, the pilot tests were conducted independently as each group was prepared to start Notably, the industrial partners invested considerable effort and resources in the test, with high-level executives and their logistics teams participating in numerous conferences.

The team held 79 calls to discuss preparations for the pilot test, ultimately deciding to offer industrial partners a token payment of $3,000 as a gesture of appreciation for their contributions However, this amount falls significantly short of compensating for the staff time invested by the partners, highlighting their strong support for the concept The participation dates for each company are noted in parentheses next to their names.

Group 1: Foot Locker and New Deal Logistics (October 2-November 14, 2009): Foot Locker operates numerous stores in Manhattan New Deal Logistics, Foot Locker‘s logistic provider, has long pioneered environmental initiatives Eight Foot Locker stores and New Deal Logistics participated

Group 2: Sysco and a Sample of Customers (December 21-January 23, 2010): Sysco delivers food products to hundreds of food stores in New York City Thirteen stores successfully completed the test, five participated partially and dropped out for reasons unrelated to the project, and another three agreed to participate but did not order products from the vendor during the pilot test (it may be that they changed vendors)

Group 3: Whole Foods Market and its Vendors (December 28-January 31, 2010): A large natural and organic grocery store chain with six locations in Manhattan This group included the four Whole Foods Market locations not subject to night delivery restrictions plus six of their vendors

Locations of the participating receivers are shown in Figure 19

Additional Details of the Pilot Test

Participants in the pilot test received a financial incentive of $2,000 to encourage successful involvement, which was higher than the incentives considered during the research due to the initial setup costs of transitioning from regular to off-hours Additionally, carriers were offered $300 per truck participating in the pilot to help cover their setup expenses, as they would benefit from operating during off-hours, allowing for a smaller incentive compared to that of the receivers.

The remote sensing project utilized GPS-enabled smartphones, specifically the MWg Zinc ii, which features a robust 500 MHz processor, a bright 2.8-inch touch screen, and a high-quality SiRf Star_iii GPS receiver Each device came equipped with CoPilot|Live version 8 software, allowing for seamless navigation without driver distraction, as only a single button was needed to initiate the process The smartphones logged essential data, such as GPS position and velocity, every three seconds and transmitted this information to a secure website for real-time tracking CoPilot|Live is tailored for the logistics industry, ensuring compliance with commercial truck routing restrictions while enhancing safety and operational efficiency.

During the pilot test, participating carriers received cellular smartphones for data collection Companies with existing GPS equipment were encouraged to utilize the provided smartphones but had the option to share their data with the research team instead Many participants chose this alternative, which, despite offering less frequent data updates—averaging once per minute without real-time access—allowed for the collection of background performance data from a broader fleet of trucks across the metropolitan area Additionally, passive GPS data loggers served as a backup in some instances.

PILOT TEST RESULTS: REMOTE SENSING AND OPINION SURVEYS

Sysco Base Case Conditions

To aid the research team in establishing base case conditions, Sysco provided remote location-time waypoint data from its trucks operating out of the Jersey City, New Jersey depot This data covers the period from December 1, 2009, to January 31, 2010, along with an additional six weeks from September 1 to October 31, 2009 The accompanying figure illustrates the GPS waypoint data near Manhattan, color-coded to reflect instantaneous speeds, with pure blue representing 35 mph and pure red indicating 5 mph; waypoints with speeds below 5 mph are excluded, resulting in many overlapping waypoints.

Figure 20: Geographic Display of Remotely Sensed Waypoint Data Near Manhattan

6.1.1 Overall characteristics of Sysco’s base case conditions remotely-sensed GPS position-time waypoint data

The speed distribution analysis of 3,561 customer-to-customer (c2c) tour segments in Manhattan reveals that travel is notably slow, with only 1.85% of segments exceeding an average speed of 10 mph and a median speed of just 3.7 mph In contrast, the median average speed for segments between non-Manhattan customers from the same Jersey City depot is significantly higher at 10.7 mph, over 2.5 times faster than that of Manhattan This stark difference highlights the challenges of urban travel within Manhattan compared to surrounding areas.

Figure 21: Cumulative Speed Distribution of the 3,561 Manhattan c2c Tour Segments

Figure 22: Cumulative Speed Distribution of the 32,272 Non-Manhattan c2c Tour Segments

Figure 23: Comparison of c2c AverageSpeed for Manhattan and Non-Manhattan Customers

6.1.2 Tour segment speeds by time-of-day (ToD)

6.1.2.1 Manhattan customer-to-Manhattan customer tour segment AverageSpeed by time of day (ToD), Sysco base case conditions

The congestion faced by Sysco drivers in Manhattan varies throughout the day, as illustrated in Figure 24 with box and whisker plots depicting AverageSpeeds Notably, speeds before 8 AM are significantly higher, with median values being twice those recorded after this time Unfortunately, Sysco does not operate during the early morning hours between 11 PM and 6 AM Interestingly, the median speeds show minimal variation from 8 AM to 10 PM, ranging from 3.8 mph at 8 AM to 2.4 mph at 10 PM There is a steady decline in AverageSpeed from 8 AM to 1 PM, dropping by 26% from 3.8 mph to 2.8 mph, after which speeds stabilize within a 10% range until a further 20% decrease occurs at 10 PM The hourly data exhibits tight variance within a low-speed range, and post-10 PM, less than 25% of average speeds exceed 5 mph Although the graph indicates numerous outliers at higher speeds, they are relatively few compared to the total of 3,561 data points, with fewer than 100 classified as outliers.

Figure 24: Manhattan c2c Tour Segment AverageSpeed by ToD for Sysco’s Base Case

Figure 25: Diagram Defining Elements of Box and Whisker Chart (from http://www.qimacros.com/qiwizard/Box-and-Whisker-analysis.jpg)

6.1.2.2 First and last tour segment AverageSpeed by time of day, Sysco base case

The tour begins and ends with a crossing of the Hudson River, utilizing the Lincoln Tunnel The initial segment takes drivers from Sysco's Jersey City depot to a customer in Manhattan, navigating potential congestion near the tunnel The final segment involves returning from the last customer in Manhattan back to Sysco's Jersey City facility, again relying on the Lincoln Tunnel for the crossing.

The analysis of Manhattan congestion, particularly around the Tunnel in Astonia, reveals significant differences in Average Speeds between inbound and outbound traffic The first tour segment shows a median speed of 11.8 mph, while the last segment drops to 8.3 mph, indicating a 42% decrease in speed when exiting Manhattan Although median speeds remain relatively stable throughout the day, early morning hours (3 AM and 4 AM) exhibit higher speeds due to lower congestion, possibly linked to maintenance activities at the Lincoln Tunnel Notably, speeds exceeding 15 mph are less frequent than 25% of the time until 11 AM, with the noon hour showing considerable congestion, though data reliability is questionable due to a limited sample size Additionally, no data was collected for the afternoon, evening, or very early morning hours as Sysco did not dispatch tours during those times.

During the return to the depot, the Last Leg and median Average Speeds exhibited minimal variation, with a peak speed of 9.2 mph recorded in the early afternoon at 2 PM and a low of 7.7 mph at 8 PM Average Speeds exceeded 11 mph for less than 25% of the time, and notably, no trucks returned to the depot during the morning hours.

6.1.2.3 Customer service time by time of day, Sysco base case

A recent investigation analyzed the time spent at customer locations in Manhattan using remotely sensed waypoint data Sysco conducted a total of 5,569 customer service stops, revealing a median stop duration of 1.31 hours Notably, 75% of these service stops lasted less than 2.32 hours, while fewer than 25% required less than 1 hour.

Delivery vehicles often remain stationary for extended periods at customer locations, with mid-day stops averaging 1.6 to 1.8 hours, significantly longer than early morning or late evening stops, which average around 1 hour Unfortunately, the data lacked attributes indicating service facilities, such as loading docks or parking options, and could not classify customer stops by size However, it is reasonable to assume that delivery size does not correlate with time of day The absence of pedestrian and street traffic congestion during early and late hours contributes to a notable 40% reduction in customer service time in the late evening compared to late morning An analysis of customer service times in specific Manhattan areas, including Lower Manhattan, Rockefeller Center, and Times Square, further highlights these trends.

Box and whisker charts illustrate customer service times for specific Midtown locations throughout the day, revealing that service times peak during late morning and early afternoon compared to late evening and early morning When analyzing the data for Times Square and Midtown together, it becomes evident that the median service time in the late morning is significantly higher, being 103% or 53 minutes longer than the overnight period Additionally, in the afternoon, the median service time increases by 34%, or 17 minutes longer than during the overnight hours.

Figure 26: First Leg AverageSpeed by ToD for Sysco Base Case

Figure 27: Last Leg AverageSpeed by ToD for Sysco Base Case

Figure 28: Manhattan Customer Service Times by ToD for Sysco Base Case

Figure 29: Lower Manhattan Customer Service Times by ToD for Sysco Base Case

Figure 30: Rockefeller Center Customer Service Times by ToD for Sysco Base Case

Figure 31: Times Square Customer Service Times by ToD for Sysco Base Case

Figure 32: Midtown Customer Service Times by ToD for Sysco Base Case

Figure 33: Midtown + Times Square Customer Service Times by ToD for Sysco Base Case

Figure 34: Manhattan Customer Service Times by Major ToD Segments for Sysco Base Case

The findings reveal significant implications regarding delivery efficiency in Manhattan, where congestion impacts travel speeds, averaging under 3 mph Sysco's analysis showed that 42.8% of customer service stops occurred during late morning, consuming 5,265 hours, or an average of 2.21 hours per stop If these stops were shifted to overnight hours, utilizing an average service time of 1.19 hours, Sysco could have saved 2,439 hours, equating to a 61-minute reduction per stop This adjustment could potentially enable tours to serve 50% more customers, highlighting a substantial productivity opportunity Overall, transitioning to overnight deliveries could reduce total truck tour time by 16.5% for Sysco's Manhattan customers, suggesting that improved efficiency in customer service may yield significant benefits beyond just mitigating congestion.

Foot Locker/New Deal Logistics Pilot Test

New Deal Logistics (NDL) offers a comprehensive range of goods movement services in the northeast, specializing in less-than-truckload (LTL) trucking from its Kearney, New Jersey depot The company delivers products to various retailers, including Foot Locker stores in Manhattan, which participated in a pilot test requiring off-hour deliveries During this pilot, NDL utilized four GPS smartphones to monitor the operational characteristics of its trucks serving Manhattan customers and shared GPS location time data collected through its fleet management activities with the research team.

6.2.1 The remotely-sensed GPS position-time data

6.2.1.1 NDL remotely-sensed GPS fleet management position-time data

NDL shared its fleet management position-time data for seven trucks servicing Manhattan customers during the week of November 4 to November 11.

In 2009, NDL's fleet management data included key metrics such as a unique truck identifier, date, time (local EST), ignition status, location reference, GPS-derived speed, and GPS coordinates (latitude and longitude) Data is recorded every two minutes while the truck is in motion, ensuring comprehensive tracking of each vehicle's performance and location.

92 arrival (―stop‖), departure (―start‖) and ignition on/off When stopped at the depot or at a customer location, data are recorded much less frequently, typically every 45 minutes

Ignition Status is not a reliable indicator of a truck's arrival or departure, as ignition-off messages may not always reach the fleet management system due to communication dead zones Therefore, while an ignition-off message suggests a stop, it is insufficient on its own A more effective approach involves calculating the distance and average speed between sequential locations; if the distance is minimal and the speed is below 3 mph over a span of at least 6 minutes, a stop is confirmed This algorithm effectively identifies stops, except in cases of severe congestion, such as in the Lincoln or Holland tunnels, where crossing times may exceed 6 minutes and speeds may drop below 3 mph However, the algorithm discards these instances due to the clear identification of GPS points at both tunnel entrances No other significant situations were found where GPS loss or communication issues led to false positives.

Table 10: Sample of NDL Remotely-sensed GPS Position-time Data

NDL Remotely Sensed GPS Vehicle Location Data

Driver_ID date-Time Ignition

Status NDL Location Reference Speed

500 Torres M 11/4/2009 21:42 On 8 E 58th St NY 27 40.76324 -73.973191

500 Torres M 11/4/2009 21:44 On 63 W 57th St NY 24 40.76416 -73.976462

500 Torres M 11/4/2009 21:46 On 971 8th Ave NY 6 40.76686 -73.982933

500 Torres M 11/4/2009 21:48 On 485 W 56th St NY 2 40.76857 -73.988338

500 Torres M 11/4/2009 21:50 On 523 W 57th St NY 2 40.76949 -73.989262

500 Torres M 11/4/2009 21:52 On 569 W 57th St NY 4 40.7702 -73.990258

500 Torres M 11/4/2009 21:54 On 756 11th Ave NY 34 40.768 -73.993031

500 Torres M 11/4/2009 21:56 On 624 11th Ave NY 19 40.76316 -73.996516

500 Torres M 11/5/2009 7:16 On 202 Pennsylvania Ave NJ N/A 40.73906 -74.101404

Figure 35 illustrates the 40 tours conducted by NDL in Manhattan during the week of November 4, 2009, using GPS data from NDL's fleet management system Each tour segment is distinctly colored, highlighting the routes taken The Kearney, NJ depot is marked, showing that the main access point to Manhattan is through the Lincoln Tunnel, as commercial vehicles operated by NDL were restricted from using the Holland Tunnel during the pilot test.

A close visual inspection of the GPS-derived position data, as illustrated in Figure 36, revealed high-quality results with no signs of adverse urban canyon effects The data points between stops followed feasible routes and showed minimal scatter, with tight clustering around customer arrival and departure locations, exhibiting only minor deviations of tens of meters Consequently, it can be concluded that the GPS location data is of exceptional quality, supporting accurate computations of travel distances and average speeds with a precision of 0.1 mph.

Figure 35: Cartographic Display of All of the NDL Position-Time Data

(7 trucks during the week of Nov 4, 2009)

Figure 36: Cartographic Display of NDL Remotely Sensed GPS Vehicle Position-Time Data

(showing urban canyon scattering that occurs around midtown Manhattan)

The analysis of urban tour segments reveals that instantaneous speed fluctuates significantly due to frequent stops and traffic conditions, making it challenging to accurately characterize average speed with limited observations Data collected every two minutes is insufficient for segments that are typically short in distance However, reliable location and time data for each tour segment allows for the calculation of an average customer-to-customer (c2c) speed This speed is determined by dividing the computed distance traveled by the measured travel time, using the great circle distance between sequential intermediate locations The travel time is calculated from the difference between the arrival and departure timestamps Although there may be slight underestimations in travel distance due to zigzagging routes, the discrepancies are minimal, consistently under 3% for sampled routes Additionally, average speeds are calculated for the initial and final segments of each tour, from the depot to the first customer and back to the depot from the last customer.

The raw NDL data were analyzed to create 40 detailed summaries of tour and trip segments, facilitating graphical representation Each tour starts and concludes at NDL's depot in Kearney, NJ, although there are rare instances where vehicle location data ends before returning to the depot, as illustrated by a vehicle's last recorded position at the Lincoln Tunnel Between the initial departure and final return segments, individual customer-to-customer segments are identified, with nearby landmarks assigned to enhance cognitive relevance while ensuring customer privacy Key attributes for each segment, as shown in the data, include drive time in hours, drive distance in miles, average speed in mph, and time to the next trip (T2NxtTrp), which reflects customer service time for all but the last segment The last segment's T2NxtTrp indicates the time spent at the depot before the next trip, though it often results in a large negative value that is typically disregarded.

Individual tours can be characterized by several key attributes, including trip departure date and time, the number of customer stops, tour length, tour duration, drive duration, stop duration, and average drive speed A visual representation of typical tour segments from 40 NDL Manhattan pilot test tours is illustrated in Figures 37 to 39 For instance, Figure 37 highlights the tours for vehicle_ID #500 during the week of November 4, 2009, with each segment color-coded for clarity Figure 38 zooms in on a specific segment from the depot to Manhattan on November 11, 2009, showcasing a circuitous route taken to bypass heavy congestion at the Lincoln Tunnel, which added nearly 1 mile to the trip length This 13.6-mile journey took 0.93 hours, resulting in an average speed of 14.5 mph; however, no data is available to ascertain the time saved or lost due to this detour.

Table 11: Sample NDL Tour Data

Veh_ID Trip_ID Obs City State City State Lon Lat Lon Lat

500 1 14 Kearney NJ Union City NJ -74104889 40739200 -74023182 40764800

500 2 12 Union City NJ New York NY -74025031 40763020 -73991964 40735080

500 3 2 New York NY New York NY -73992249 40734860 -73996231 40736140

500 4 11 New York NY Triborough NY -73996089 40737000 -73955271 40779660

500 13 6 Midtown NY Times Square NY -73987911 40749940 -73991680 40755700

500 14 3 Times Square NY Times Square NY -73991538 40755130 -73987911 40753420

Veh_ID Trip_ID Obs Date Time Date Time Hours Miles MPH T2NxtTrp

In Figure 39, a creative route is illustrated for NDL's Kearney depot to the Lincoln Tunnel, where vehicle_ID #509 skillfully navigated exit 15x of the NJ Turnpike and Secaucus Road to bypass significant congestion near exit 16 and the NJ 495 viaduct Departing the depot at 8:15 AM, the truck successfully exited the Lincoln Tunnel just 45 minutes later, arriving at its first customer by 9:28 AM, achieving an average speed of 17.0 mph.

All 40 tours included at least one customer in Manhattan, necessitating travel through its streets and avenues Each tour featured two water crossings, primarily utilizing the Lincoln Tunnel, while some opted for the George Washington Bridge Additionally, a few tours crossed the East River via the Verrazano Narrows Bridge and Staten Island to accommodate customers in Brooklyn and Long Island.

Figure 37: NDL Manhattan Tours of Vehicle_ID #500 for the Week of Nov 4, 2009

Figure 38: NDL Depot-to-Manhattan Tour Segment for Vehicle #500 on Nov 11, 2009

Figure 39: NDL Depot-to-Manhattan Tour Segment for Vehicle #509 on Nov 10, 2009

The tours exhibited significant variability in customer numbers, with a median of 5 customers served across 40 tours The accompanying box and whisker graph illustrates this variation, revealing a range that peaks at 15 customers for a single tour Detailed statistics are provided in the accompanying tables, highlighting the diverse sequences and customer counts for each tour.

6.2.2.1 How time is spent on NDL tours by type of activity

Tours for drivers primarily involve a day's work, with time spent at the depot before and after each delivery Remote sensing data captures the driver's location from departure to return, highlighting a consistent pattern in the journey from the depot to the first customer and back Notably, travel routes between customers in Manhattan show significant similarities, especially when contrasted with trips outside the area Therefore, the analysis concentrates on three key tour segments: (1) the route from the depot to a Manhattan customer, (2) travel between Manhattan customers, and (3) the return journey from a Manhattan customer to the depot Only tours servicing Manhattan customers are included in this analysis, reflecting a focus on the unique characteristics of these routes.

The analysis focuses exclusively on the Manhattan segments of the tour, specifically the depot-to-Manhattan and Manhattan-to-Manhattan routes Internal segments in New Jersey and Brooklyn are excluded to ensure a concentrated examination of the primary Manhattan connections.

Figure 40: Distribution of NDL Customers Served Per Tour during Field Test

The tour duration characteristics for NDL tours catering to Manhattan customers reveal that the median tour lasts just under 8 hours, with shorter tours exhibiting greater variability Customer service times are evenly distributed around a median of 4 hours, with the first quartile at 2 hours and the third quartile just below 6 hours Notably, the Hudson River crossings significantly extend the tour duration, averaging nearly 1.5 hours for a distance of about 25 miles Additionally, the median c2c drive time is low due to the clustering of Manhattan customers, resulting in minimal drive times for tours with fewer customers, with a median of less than half an hour.

Sysco Pilot Test

Sysco, a global leader in food distribution, serves a diverse range of customers who prepare meals away from home, with multiple distribution centers across North America Its Jersey City, NJ depot, Sysco Metro New York, LLC, caters to clients in the New York metropolitan area, primarily using class 6 box and combination trucks for daily tours that align with customer delivery schedules during normal business hours The company has initiated an off-hour Manhattan delivery tour and is eager to expand this service As part of a pre-pilot test from July 6 to August 6, 2009, Sysco collected and shared GPS position-time data for its fleet, and participated in an expanded off-hour delivery pilot from December 1, 2009, to January 31, 2010.

6.3.1 Sysco’s remotely-sensed GPS position-time data

Sysco's fleet utilizes advanced XataNet asset tracking and management, featuring precision GPS units in each truck This system records and transmits truck locations every one to two minutes during movement and every thirty minutes when stationary Additionally, time-stamped data is collected upon arrival and departure from various locations, providing comprehensive tracking capabilities.

Table 12: A Snippet of the Sysco Remotely-sensed Position-time Data

Message # VehicleID Longitude Latitude Date-Time Status Speed Heading

The longitude and latitude location data demonstrated a precision exceeding 100 meters, despite the challenging urban canyon environment of Manhattan This area is prone to significant signal scatter, particularly near tunnel entrances and on narrow streets with minimal building setbacks, such as Wall Street, Madison Avenue, and 56th Street Nevertheless, an inspection of the geographical data display reveals no evident issues arising from this adverse environment.

105 are depicted in Figure 49 through Figure 52, which suggests that there are any significant location outliers with precision worse than 30 meters (100 feet)

6.3.1.1 Computation of tour segment average speed

The study provided precise date-time data, including seconds, though not displayed in Table 12 Due to the unreliability of instantaneous speed, the focus shifted to calculating "average speed," defined as the distance traveled divided by the travel time between stops, such as depots or customer locations The distance was determined by the CoPilot software, which calculated the most probable route taken between closely spaced data points Travel time was measured as the difference between the designated start and stop points for each route segment Given the accuracy of both the computed distance and travel time, the resulting "average speed" is presented with high confidence, measured in tenths of miles per hour.

Accurate computation of average speed relies heavily on correctly identifying a vehicle's departure and arrival at customer or depot locations Position data is recorded during both movement and stops, necessitating careful distinction between actual stops and those caused by extreme congestion, such as near Hudson River crossings The vehicle status Boolean alone is not a reliable determinant for these data points Therefore, an algorithm was created to analyze preceding and subsequent data points to refine the transition timing If a vehicle arrives at a stop and the distance to the previous point is less than 0.1 miles, the arrival is adjusted to that earlier point Conversely, if the next data point is within 2 minutes and the distance exceeds 0.1 miles, the arrival at a customer or depot is moved forward Similar adjustments are applied to departure times from these locations.

The algorithm for identifying departures and arrivals was enhanced by incorporating transitions in the GPS position-time data rate, which varies significantly between moving (approximately 1 minute) and stopped (around 10 minutes) states Additionally, the algorithm required that all points within the cluster between a new arrival and departure be located within a 0.1 mile radius, with the stop duration exceeding 0.1 hour (6 minutes).

The median duration of stops is slightly less than one hour, with only a few outliers exceeding three hours It is believed that these outliers may include undetected customer stops However, their exclusion does not adversely impact the evaluation of congestion experienced by the vehicles.

The analysis identified 106 instances that could have enhanced the customer-to-customer average speed dataset; however, distinguishing between stops and severe congestion, particularly near the Hudson River crossings, proved challenging Fortunately, Sysco's customer base is not located near these congested areas, allowing for the consolidation of stops in those regions This approach effectively segmented the Sysco pre-pilot test data into 1,766 tours, which included 14,929 service stops at customer locations, culminating in a total of 16,695 tour segments.

Figure 48: Box and Whisker and Sorted Order Graphs of Duration of Customer Stops

6.3.1.2 Cartographic display of Sysco GPS remotely-sensed position data

The remotely-sensed pre-pilot test GPS position data shared by Sysco with the research team reveals significant activity in Manhattan and its surrounding areas, as illustrated in Figure 49 Out of 1,766 tours originating from Sysco's Jersey City depot, 471 are dedicated to serving customers in Manhattan, encompassing a total of 2,613 customer-to-customer (c2c), depot-to-customer (d2c), and customer-to-depot (d2c) tour segments The average speeds of these segments help quantify the congestion faced by the goods movement industry in Manhattan, particularly during different times of day Figure 50 highlights a tour segment from the Jersey City depot to a customer on the Upper West Side, showing a clear alignment between the location data and the expected travel path Notably, a cluster of points near the Lincoln Tunnel entrance and exit indicates potential customer locations that were excluded by the stop insertion algorithm In Figure 51, data typical of shorter c2c segments in Manhattan shows minimal scatter, while Figure 52 presents a long tour segment from Sysco's depot, further illustrating the patterns of goods movement in the area.

Traveling from Jersey City to a customer location in Brooklyn may seem short in distance, but the optimal route for drivers is crucial During this trip, inbound commercial traffic was prohibited from using the Holland Tunnel, a restriction that has recently been lifted for light-duty commercial vehicles However, this ban still applies to combination (truck-tractor) class 6 and class 8 commercial vehicles, which make up approximately half of Sysco's fleet.

Figure 49: Sysco Tour Segments Contained in the Pre-Field Test Dataset

Figure 50: Sysco Tour Segment from Depot to an Upper West Side Customer

Figure 51: Sysco Manhattan Customer-to-Customer Route Segment

Figure 52: Sysco Depot to Brooklyn Customer Segment

The raw Sysco data were processed to create summaries of individual tours and trip segments, as illustrated in Table 13 Each tour starts and ends at Sysco's depot in Jersey City, featuring segments between customer stops, which are identified by nearby landmarks to ensure cognitive relevance while protecting customer privacy The data includes essential details such as the date and time each segment concludes, along with summary metrics for each trip segment, including drive time, segment length, average speed (calculated as length divided by time in miles per hour), and the duration spent serving each customer.

The T2NxTr metric for the final tour segment represents the duration a vehicle remains at the depot before being put back into service It is important to note that the value for the last tour of a specific truck should be excluded, as it reflects the idle time between the conclusion of that tour and the commencement of the first tour recorded in the database for the subsequent truck_ID.

Table 13: Sample of Sysco Pre-field Test Tour Summary Data

Origin Destination Origin Destination Origin

Veh_ID Trip_ID Observ City State City State Lon Lat Lon Lat Date Time

Trip_ID Date Time Hours Miles MPH T2NxTr FrMan FrNJ ToMan toNJ ToD

Historically, customers in Manhattan have favored service during standard business hours, a trend evident in Sysco's operational data An analysis of the depot departure times for 438 tours servicing Manhattan customers reveals that only 50 departures took place before 7 AM, with the majority occurring before 6 AM This pattern likely reflects an effort to circumvent the significant congestion typically encountered during peak hours.

110 the Lincoln Tunnel around 7 AM Departure rates remain relatively constant from 7 AM through 9 AM when 70% of the departures occur Less than 50 departures occur between 10 PM and 4 AM

Figure 53: Depot Departure Time: Sysco Tours Serving Manhattan Customers

The return to the depot is primarily concentrated in the evening, with most returns occurring between 5 PM and 9 PM, while fewer than 20% of tours return after 10 PM Notably, very few returns happen before 1 PM, indicating a clear pattern in the timing of depot returns.

Figure 54: Depot Return Time: Sysco Tours Serving Manhattan Customers

The distribution of customers served per tour from Sysco's Jersey City depot reveals that the median is approximately 9 customers across 1,766 tours The data shows a range of low outlier values, with some tours serving only 1 or 2 customers, while others serve as many as 15, 16, or 17 customers Notably, 50% of the tours serve between 7 to 10 customers A similar pattern is observed in the tours conducted by Baldor, a vendor for Whole Foods Market.

Figure 55: Distribution of Customers Served Per Tour in Sysco’s GPS Data

6.3.2.2 Manhattan tour segment length characteristics

Whole Foods Market/Baldor Specialty Foods Pilot Test

Baldor Specialty Foods, located in the Hunts Point Market of Bronx, NY, is a leading food service company that focuses on sourcing and distributing specialty food items Utilizing a fleet of 150 refrigerated box trucks, Baldor delivers to establishments across the New York metropolitan area and as far as Boston and Philadelphia Recently, the company conducted a pilot test for off-hour early morning deliveries to Whole Foods Market stores in Manhattan, integrating these deliveries into standard operational tours that service multiple customers These delivery routes are designed to optimize driver productivity by maximizing drive time within the constraints of the workday, rather than being limited by truck capacity.

To assess the effectiveness of off-hour deliveries, Baldor provided the research team with remotely-sensed GPS position-time data for its trucks making early morning deliveries to Whole Foods Market in Manhattan, as well as other trucks servicing customers in the New York metropolitan area This data collection occurred over a two-month period from December 1, 2009, to January 30, 2010, during which data was available for 51 days The study highlights the operational dynamics of Baldor's delivery business.

Tours consistently depart and return to Hunts Point, serving a clientele primarily from Manhattan; however, each day presents unique variations in the customer list, delivery volumes, and the sequence of deliveries Notably, significant quantifiable performance differences are observed at various times throughout the day.

6.4.1 The remotely-sensed GPS position-time data

Table 14 presents a summary of Baldor's GPS position-time data for trucks, featuring unique identifiers for each vehicle and driver, along with date, time (EST), GPS-derived latitude and longitude, speed, and arrival/departure attributes Data is recorded every two minutes while the truck is in motion, as well as during key events such as arrival ("stop") and departure ("start") The reliability of capturing position-time data at customer locations has been confirmed, with less frequent recordings (approximately every 45 minutes) when stationary A visual inspection of the geographic data, illustrated in Figure 65, indicates high-quality GPS data without significant urban canyon effects The recorded points between stops followed feasible routes with minimal scatter, and those around arrival and departure locations were closely clustered, showing only minor deviations of a few meters This confirms that the GPS location data is of excellent quality, supporting accurate calculations of travel distances and average speeds within 0.1 mph.

The analysis did not utilize the provided speed data due to the significant fluctuations in instantaneous speed during trips, which include frequent starts, stops, and varying durations of intermediate stops caused by traffic control and congestion Accurately characterizing the average speed of these highly variable conditions necessitates a much higher data collection rate, as a two-minute interval is inadequate for capturing the average speed of urban tour segments that are usually just a few miles long.

Table 14: Sample of Baldor Remotely-sense GPS Position-time Data

Local Time Baldor Location Reference City St

1120 1/2/2010 4:14:19 BALDOR FCD Bronx NY 0 40.8097224 -73.8734841 Ign on

1120 1/2/2010 4:14:55 BALDOR FCD Bronx NY 2 40.8097224 -73.8734841 Start

1120 1/2/2010 4:16:55 TEAMSTERS LOCAL 202 Bronx NY 0 40.8127932 -73.8811312 Moving

1120 1/2/2010 4:18:55 SNOW FRESH Bronx NY 42 40.8118407 -73.8886139 Moving

1120 1/2/2010 4:20:55 1056 Leggett Ave Bronx NY 0 40.8129731 -73.8977703 Moving

1120 1/2/2010 4:22:55 407 Bruckner Blvd Bronx NY 40 40.8080417 -73.9054737 Moving

1120 1/2/2010 4:24:55 134 Bruckner Blvd Bronx NY 31 40.8040029 -73.9206641 Moving

1120 1/2/2010 4:26:55 232 E 128th St New York NY 34 40.8053903 -73.933657 Moving

1120 1/2/2010 4:28:55 142 E 125th St New York NY 9 40.8043631 -73.9373676 Moving

1120 1/2/2010 4:30:55 1783 Lexington Ave New York NY 20 40.7951774 -73.9441796 Moving

1120 1/2/2010 4:32:55 1439 Lexington Ave New York NY 27 40.7844038 -73.9520606 Moving

1120 1/2/2010 4:34:55 1073 Lexington Ave New York NY 27 40.7729431 -73.9604192 Moving

1120 1/2/2010 4:36:55 727 Lexington Ave New York NY 27 40.7617576 -73.9685822 Moving

1120 1/2/2010 4:38:55 279 Lexington Ave New York NY 31 40.7478951 -73.9786505 Moving

1120 1/2/2010 4:40:55 146 E 33rd St New York NY 13 40.7456937 -73.9800413 Moving

1120 1/2/2010 4:42:55 166 E 23rd St New York NY 0 40.7391655 -73.9839234 Moving

1120 1/2/2010 4:44:55 168 E 17th St New York NY 11 40.7355154 -73.9864261 Moving

1120 1/2/2010 4:46:55 20 Union Sq E New York NY 2 40.7361058 -73.9894047 Moving

1120 1/2/2010 4:48:55 WHOLE FOODS UNION SQ New York NY 0 40.7346378 -73.9919725 Moving

1120 1/2/2010 4:50:43 WHOLE FOODS UNION SQ New York NY 0 40.7346378 -73.9919725 Stopped

1120 1/2/2010 4:51:13 WHOLE FOODS UNION SQ New York NY 0 40.7346378 -73.9919725 Idling

1120 1/2/2010 5:00:25 WHOLE FOODS UNION SQ New York NY 0 40.7346378 -73.9919725 Start

1120 1/2/2010 5:02:25 240 E 14th St New York NY 11 40.7330651 -73.9862161 Moving

1120 1/2/2010 5:04:17 307 E 12th St New York NY 0 40.7309281 -73.9857793 Ign off

1120 1/2/2010 5:04:18 307 E 12th St New York NY 0 40.7309281 -73.9857793 Stopped

1120 1/2/2010 5:24:37 307 E 12th St New York NY 0 40.7309281 -73.9857793 Ign on

1120 1/2/2010 5:25:39 307 E 12th St New York NY 0 40.7309281 -73.9857793 Start

1120 1/2/2010 5:26:11 222 1st Ave New York NY 0 40.7309259 -73.9829567 Ign off

1120 1/2/2010 5:26:12 222 1st Ave New York NY 0 40.7309259 -73.9829567 Stopped

1120 1/2/2010 5:38:23 222 1st Ave New York NY 0 40.7309259 -73.9829567 Ign on

1120 1/2/2010 5:39:46 222 1st Ave New York NY 13 40.7309259 -73.9829567 Start

The data provides accurate and reliable timestamps and locations for the beginning and end of each tour segment Drivers are trained to indicate departures and arrivals at customer locations with a simple tap, which the vehicle information system logs with GPS coordinates and timestamps Additionally, the system automatically records GPS data when the ignition is turned on and off as a backup To ensure accuracy, the distance between each logged data point is calculated algorithmically, allowing for precise determination of customer arrival and departure times and locations.

In a cluster of time-sequenced points, specific criteria are used to identify significant events: the distance between points must be less than 0.1 miles, and there should be a combination of vehicle statuses such as "Start" and "Stopped," or "Ignition off" and "Ignition on." The earliest point in this cluster represents the arrival time and location for the target customer, while the latest point indicates the departure time and location.

The average customer-to-customer (c2c) speed for each tour segment is calculated by dividing the total great circle distance traveled from the previous customer's departure location to the current customer's arrival location by the measured travel time This distance is determined by summing the great circle distances between sequential intermediate locations within the tour segment Measured travel time is simply the difference between the arrival and departure timestamps of the respective customers Although trips may zigzag between recorded locations, the resulting underestimation of travel distance is minimal, as the difference between great circle distances and street network path distances is consistently less than 3% in random samples.

An average speed was also computed for each tour‘s first and last segments, from the depot to the first customer and from the last customer back to the depot

Figure 65: Cartographic Display of Baldor Vehicle Position Data

The raw Baldor data were analyzed to create summaries of individual tours and trip segments, which were then prepared for graphical representation Each tour starts and concludes at the Hunts Point depot, featuring various segments between customer stops, each named after a nearby landmark to ensure relevance while protecting customer privacy The data includes the date and time marking the end of each segment, along with summary statistics for each trip segment, such as drive time, segment length, average speed (calculated as length divided by time in miles per hour), and the duration spent serving each customer (Time2NxtTrp).

Table 15: Sample Baldor Tour Data

City State City State Lon Lat Lon Lat

931 1 Start 1 12 HUNTS POINT NY Floral Park NY -73875446 40809374 -73756081 40749209

931 2 1 19 Horace Harding NY Brentwood NY -73754278 40749730 -73265081 40806290

931 7 1 12 Holbrook NY Central Islip NY -73096081 40789880 -73185015 40811048

931 8 1 18 Central Islip NY Huntington Station NY -73184042 40811036 -73417551 40886568

931 9 1 5 Huntington Station NY Huntington NY -73417551 40886568 -73426203 40870260

931 11 1 13 Northport NY Huntington Station NY -73360908 40900766 -73411978 40827874

931 12 End 1 35 Huntington Station NY HUNTS POINT NY -73411978 40827874 -73873580 40809790

Date Time Date Time Hours Miles MPH Time2 NxtTrp Seg

The data visualization in Figure 66 illustrates 51 daily Baldor tours servicing lower Manhattan from December 1, 2009, to January 31, 2010, with each color representing a different tour The analysis reveals minimal urban canyon scatter, indicating that the routes primarily utilize avenues to travel from the Hunts Point home depot to various customer locations scattered throughout Manhattan south of Central Park Key attributes of each tour include trip departure date and time, number of customer stops, total tour length, duration, drive time, stop duration, and average drive speed Figures 67 through 70 provide a cartographic display of selected typical tours from this pilot test.

On December 16, 2009, a Baldor Manhattan tour departed at 4:33:26, returning 8.05 hours later after serving 10 customers, which required 6.14 hours of service The drive time to the first customer was 0.637 hours at an average speed of 15.0 mph, with average c2c speeds of 5 mph Similarly, on December 18, 2009, another Baldor Manhattan tour left at 4:32:07 and returned 9.45 hours later after serving seven customers, needing 6.96 hours The drive time to the first customer was 0.635 hours at an average speed of 15.2 mph, with c2c speeds averaging 5.9 mph.

On January 26, 2010, the Baldor Manhattan tour departed at 4:38:25 and returned 7.98 hours later after serving 9 customers, which took a total of 5.24 hours The drive time to the first customer was 0.626 hours at an average speed of 15.2 mph, with an average c2c speed of 7.72 mph In contrast, another trip in 2010 left at 4:44:16 and returned 7.59 hours later after serving 10 customers, requiring 4.43 hours The drive time to the first customer for this trip was 0.577 hours at an average speed of 16.0 mph, with c2c speeds averaging 4.5 mph.

The tours, primarily catering to customers south of 42nd Street in Manhattan, share similarities in their early morning departures from the Hunts Point depot, with most leaving between 4 AM and early 5 AM However, they vary in details such as the number of customers served, the specific sequence of stops—starting with Whole Foods Market stores—and the congestion encountered, as indicated by each segment's average speed The variations are illustrated through standard box and whisker graphs in Figures 71 to 75, along with their corresponding tables.

Figure 66: Display of Location Data of 51 Daily Baldor Tours

Figure 67: Baldor Manhattan Tour on December 16, 2009

Figure 68: Baldor Manhattan Tour on December 18, 2009

Figure 69: Baldor Manhattan Tour on January 13, 2010

Figure 70: Baldor Manhattan Tour on January 26, 2010

Figure 71: Baldor Manhattan Customers Served Per Tour

Tours consist of various trip segments, which are categorized into two types: the segments between stops and the initial and final segments involving the home depot The home depot segments are generally longer and utilize major roadways, while the segments between customer stops typically take place on local roads.

6.4.2.1 Manhattan tour segment length characteristics

Figure 72 illustrates a significant difference in tour lengths between the initial and final segments traveling to and from the Bronx depot compared to the shorter segments between customer locations in Manhattan The Manhattan customers are closely grouped, typically within a half-mile radius, while the distance for the Bronx segments is nearly 20 times greater The median distances for these segments are approximately 10 miles, with the first segment averaging 9.3 miles and the last segment at 9.2 miles.

Figure 72: Baldor Manhattan Individual Tour Segment Lengths

Overall Results: Analysis of Pooled Data

The pilot test data were initially analyzed separately, revealing that individual data files lacked the size needed for a comprehensive analysis of time-of-day variations Consequently, the data were pooled for the final analyses discussed herein Figure 77 illustrates the travel speed distributions among Manhattan customers at various times throughout the day AverageSpeed is calculated as the ratio of actual travel distance to travel time between consecutive customers during the pilot test, with values assigned to the hour of departure from customer i The overall Median AverageSpeed was found to be 3.30 mph, with 25% of speeds falling below 2.10 mph and 25% exceeding 5.0 mph Figures 78 and 79 show AverageSpeed distributions from 4:00 to 23:00, noting that no customers were served between 23:00 and 4:00 These distributions exhibit a heavy tail, with some AverageSpeeds surpassing 15 mph, resulting in average speeds that are higher than the median The standard deviation values remain modest.

The box and whisker chart illustrates that Average Speeds in Manhattan are significantly higher during the early morning hours compared to late morning, afternoon, and early evening Specifically, the median Average Speeds in the early morning are nearly double those recorded in late morning In contrast, Average Speeds in the afternoon and early evening drop to around 3.0 mph or less, resembling walking speeds, and these low values continue into the evening It is important to note that no tours were conducted before 4 AM, leaving a gap in data regarding congestion levels during those early hours.

Figure 77: Distribution of AverageSpeed for 4,020 Individual Manhattan c2c Tour Segments

Figure 78: AverageSpeed Distributions for 5 AM to 2 PM - Manhattan c2c Tour Segments

Figure 79: AverageSpeed Distributions for 2 PM to 11 PM - Manhattan c2c Tour Segments

Figure 80: Manhattan c2c AverageSpeed by Hour of Day - Field Test (Sysco, Baldor, NDL)

Analysis of traffic data reveals that peak periods in Manhattan do not align with those in the rest of New York City, with Manhattan's traffic lagging behind Traditional traffic groupings indicate higher mean AverageSpeeds during the AM Peak and MidDay by 21.8% and 7.5%, respectively However, during Off-Hours, Manhattan's traffic shows a significant 57.4% increase in mean AverageSpeed compared to traditional groupings These findings suggest a need to redefine Off-Hours for future delivery research in Manhattan Figures 81 and 82 illustrate the differences in c2c AverageSpeeds between traditional and proposed Manhattan traffic groupings.

Figure 81: AverageSpeed by Traditional ToD Grouping

Results from the Opinion Surveys from Carriers and Receivers

After the pilot test was successfully completed, participants were requested to fill out a brief survey regarding their experiences, as detailed in the Appendix The subsequent sections present the findings from the satisfaction surveys that were returned.

Paul Cox, VP of Global Transportation & Supply Chain at Foot Locker, reported a positive experience with the OHD program, especially in larger volume stores with dedicated backroom staff, during a call on November 11, 2009 Following this, Foot Locker was contemplating the expansion of OHD to additional stores in Manhattan However, a survey conducted among store managers failed to clarify that all extra costs would be covered by the financial incentive, leading many managers to perceive the OHD program negatively due to the unexpected expenses they faced The feedback from the eight participating Foot Locker stores reflects these sentiments.

Figure 82: c2c AverageSpeed by Manhattan ToD Grouping

What was your impression of off-hour deliveries?

Very Favorable Average Response Very Unfavorable

How much did receiving off-hour deliveries affect your operations?

How were your operations affected?

 ―The store‘s shipment arrived during store closing The shipment remained on floor until the next morning causing the store to be cluttered at start of next business day.‖

Due to the late arrival of the delivery truck, the store had to reschedule associates for shipment processing, resulting in incurred overtime costs An additional associate was brought in to manage the shipment during this delayed delivery period.

 ―Store had to adjust work schedule to receive product.‖

To accommodate both parcel and pool deliveries, the store adjusted the associate's schedule, requiring them to work as a receiver all day As a result of the accumulated hours, this associate was granted an additional day off during the week Consequently, all other associates were also rescheduled to ensure coverage during the receiving associate's day off.

 ―Receiving off-hours slowed processing and start for the next business day Shipment boxes were in the way of store opening.‖

 ―Not as many associates were available to work at the start of business day It was difficult to schedule associates to receive shipments.‖

 ―Pool agent arriving after store closure Store paid overtime to keep personnel late.‖

If it were up to you, how likely are you in the future to request deliveries from your vendors in the off- hours?

Very Likely Average Response Very Unlikely

What did you like about receiving deliveries in the off-hours?

 ―Store could process freight faster in off-hours.‖

 ―Receiving shipment did not compete with customers Boxes from shipment were not on the sales floor Store was clean for early morning shoppers.‖

 ―Less traffic in store – Very few customers were interrupted due to the receiving of shipment.‖

 ―Receiving shipment did not compete with business Customers were not interrupted by receiving shipment.‖

 ―Store manager did not like off-hours deliveries.‖

What did you dislike about receiving deliveries in the off-hours?

 ―Store could not get new product out on sales floor until the next business day causing a loss of sales.‖

 ―Late hours – Clean up of store in evening plus late receiving of freight necessitated store personnel to work late adding extra overtime.‖

Store parcel freight is delivered in the early morning, necessitating the scheduling of stock personnel late in the evening to receive pool freight If the pool agent arrives late, store staff must also extend their hours, leading to overtime payments for associates to handle the shipment.

 ―Changing of schedules for all personnel to accommodate late pool deliveries.‖

 ―Do not like receiving shipment so close to store closing time.‖

 ―Missed sales opportunities – Large shipments could not be processed until the next morning Parcel shipments also received in the morning got mixed with pool shipments.‖

 ―Manager had to stay late after regular store hours to receive freight Store had to pay overtime to associates to receive shipment.‖

Responses indicate that receivers generally disapprove of off-hour deliveries (OHD), citing challenges related to scheduling and operational disruptions Key concerns include the hassle and cost of rescheduling staff, as well as delays caused by shipments obstructing business operations the following day While nearly half of the receivers expressed dissatisfaction with OHD, the other half acknowledged benefits such as faster freight processing and minimal interruptions to customer service Many criticisms stem from the pilot test's temporary nature, suggesting that scheduling issues and extra costs may diminish over time with proper incentives Ultimately, while OHD could enhance efficiency for receivers, convincing them to embrace this approach may require addressing their initial resistance to rescheduling.

The project team analyzed satisfaction surveys from twelve Sysco receivers, revealing an average overall impression score of 1.50 on a scale from one to five, where "1" indicates "Very Favorable" and "2" signifies "Favorable." When asked about the likelihood of requesting OHD in the future, nine respondents indicated they were "Very Likely," one was "Likely," and the remaining two were uncertain Additionally, six participants utilized unassisted deliveries during the pilot, with five expressing interest in continuing this service during off-hours, contingent upon resolving liability concerns.

The implementation of Off-Hours Deliveries (OHD) significantly improved operational efficiency for receivers, leading to positive outcomes such as easier and more reliable deliveries that consistently arrived on time Although there were minor additional costs associated with rescheduling and handling returns or exchanges, the overall benefits included cost savings from pre-business hour orders and reduced labor as deliveries did not require manual handling As a result, nearly all receivers expressed a strong likelihood of requesting off-hours deliveries in the future The key advantages highlighted were enhanced reliability, convenience, and minimal disruption to business operations, while the only notable drawbacks were the extra costs, delayed verification, potential return issues, and security concerns This summary reflects the feedback from Sysco's participating customers.

What was your impression of off-hour deliveries?

Very Favorable Average Response Very Unfavorable

How much did receiving off-hour deliveries affect your operations?

No Changes Average Response Drastic Changes

How were your operations affected?

 ―Changes in operations dealing with delivery.‖

Establishing order before launching a business is essential for success Receiving deliveries during peak times can be chaotic, but having everything organized and stored away before opening significantly eases the process This approach also allows for a more efficient strategy in managing orders and stocking inventory.

 ―Required additional employee scheduling and costs, required additional time for returns/exchanges of items.‖

 ―They were always on time and we save by not having to put everything in the walk-in ourselves.‖

 ―We prep a lot of food for stores and delays in delivery time will cost us real labor dollars.‖

 ―It was much better than having to interrupt our customer space with boxes and mess.‖

If it were up to you, how likely are you in the future to request deliveries from your vendors in the off- hours?

Very Likely Average Response Very Unlikely

What did you like about receiving deliveries in the off-hours?

 ―Receiving deliveries in the off-hours will not interrupt the operation during the busy hours.‖

Establishing order before starting a business is crucial, especially when deliveries coincide with busy periods Having inventory organized and stored away prior to opening enhances efficiency and allows for improved inventory management and ordering practices.

 ―Receiving deliveries during the off-hours is more convenient and plus I receive my order earlier in the day.‖

 ―There was no/little disruption of customers overlook experience due to deliveries, quick check in and rotation into stock, did not affect the kitchen cooking process.‖

 ―They were never late for delivery!!‖

 ―Reliability and they put stuff away.‖

 ―It was much better than having to interrupt our customer space with boxes and mess.‖

 ―The advance receipt of goods.‖

What did you dislike about receiving deliveries in the off-hours?

 ―The additional cost of assigning a staff to receive the orders in the off-hours.‖

To ensure the safe storage of products before staff arrival, we coordinated access for the driver Although we did not encounter any issues in this instance, larger-scale operations could potentially lead to challenges that we would need to address.

 ―Additional cost for employee dedicated to handle delivery, return/exchange process is made more difficult.‖

 ―If you are shorted or get wrong item, you cannot give to the driver the next day Sales guy has to come and get it.‖

 ―It is problematic to return items.‖

 ―Can't check quality or wrong items.‖

If all liability issues were addressed, would you be interested in receiving unassisted deliveries (e.g driver places goods in a secure location at your establishment)?

Very Interested Average Response Very Uninterested

The project team collected satisfaction surveys from all four Whole Foods Market locations, revealing an average overall impression rating of

In a recent pilot test involving Whole Foods Market stores, deliveries were conducted during off-hours, resulting in minimal disruption to operations The feedback from participating stores indicated a lack of interest in the concept of unassisted deliveries, with survey results reflecting a favorable rating of 2 and a neutral rating of 3.

What was your impression of off-hour deliveries?

Very Favorable Average Response Very Unfavorable

How much did receiving off-hour deliveries affect your operations?

No Changes Average Response Drastic Changes

How were your operations affected?

 ―Most of our deliveries were already overnight.‖

 ―Scheduling of deliveries, more timely deliveries due to less traffic at night.‖

If it were up to you, how likely are you in the future to request deliveries from your vendors in the off- hours?

Very Likely Average Response Very Unlikely

What did you like about receiving deliveries in the off-hours?

 ―Less traffic on the street.‖

 ―Avoiding traffic from rest of building.‖

 ―Less congestion outside receiving Less late deliveries.‖

 ―Less traffic congestion and pedestrians.‖

What did you dislike about receiving deliveries in the off-hours?

If all liability issues were addressed, would you be interested in receiving unassisted deliveries (e.g driver places goods in a secure location at your establishment)?

Very Interested Average Response Very Uninterested

The project team analyzed satisfaction surveys from eight participating carriers regarding their views on OHD Among them, two carriers expressed a "Very Favorable" opinion, four had a "Favorable" view, one remained "Neutral," and one carrier reported a "Very Unfavorable" perspective, citing issues with extended wait times and expressing reluctance to perform OHD in the future In contrast, the other seven vendors indicated a strong willingness to accommodate off-hour deliveries, stating they would be "Very Likely" or "Likely" to fulfill such requests from customers.

Managers reported that OHD has effectively reduced parking violations and enhanced parking availability However, they expressed concerns regarding driver safety and the inconvenience of waiting for customers to open They suggested that implementing unattended deliveries could address this issue The following is a summary of the feedback gathered from the surveys completed by the managers of the participating carriers.

What was your impression of off-hour deliveries?

Very Favorable Average Response Very Unfavorable

How much did receiving off-hour deliveries affect your operations?

No Changes Average Response Drastic Changes

How did making off-hour deliveries affect your costs?

Moderate Decrease Average Response Moderate Increase

If it were up to you, how likely are you to make deliveries the off-hours if requested from your customers?

Very Likely Average Response Very Unlikely

What did you like about receiving deliveries in the off-hours?

 ―Reduced parking violations, available parking for our vehicles, fewer traffic delays.‖

TRAFFIC SIMULATION

Introduction

An off-hour delivery (OHD) program aims to enhance highway network efficiency by reducing the number of trucks and commercial vehicles during peak congestion times, thereby improving highway speeds and decreasing travel times However, accurately estimating its effects poses challenges due to the complexity and data requirements of freight planning models Many city transportation agencies and metropolitan planning organizations have already created traffic planning and simulation models This article details the application of the Best Practice Model (BPM), developed by the New York Metropolitan Transportation Council (NYMTC), alongside a mesoscopic sub-simulation model to assess the impact of the OHD program on Manhattan's traffic network A research methodology is established to adjust and distribute freight trip tables based on the participation rate of freight vehicles in the OHD program, with necessary calibrations made to the models for accurate analysis.

After obtaining the models' results, a thorough analysis is performed to assess the effectiveness and impacts of the modeled scenarios, particularly focusing on predicted changes to traffic networks, including travel times and speeds Additionally, economic evaluations are conducted to quantify savings from reduced congestion and other external benefits The results from both models are compared and analyzed, leading to a discussion on their applications Finally, the analysis explores the integration of the OHD program with congestion pricing and dynamic pricing initiatives.

Traffic Modeling Tools Evaluated

To develop a methodology for assessing the traffic impacts of freight congestion control measures, a review of existing literature in freight planning was conducted This review emphasized freight planning modeling and the utilization of available traffic planning and simulation tools to evaluate the effects of policy measures The subsequent sections summarize these key topics.

While significant research has focused on quantifying the traffic impacts of congestion mitigation programs, there is a noticeable lack of studies specifically addressing initiatives for trucks and commercial vehicles Current programs aimed at improving highway freight transportation are still in their experimental stages, with only a few cities having implemented effective congestion control measures It is well-documented that truck and commercial vehicle traffic adversely affects highway networks, contributing to overall congestion and inefficiencies.

Truck traffic contributes to congestion on highways and local roads due to their larger size and slower speeds compared to automobiles Research indicates that the presence of trucks significantly reduces the flow rate of traffic, exacerbating congestion in areas with high vehicle volumes (Ioannou et al., 2001).

Freight traffic from delivery vehicles significantly contributes to urban congestion, exacerbated by double-parking and blockages due to insufficient parking during peak delivery hours (Holguin-Veras et al., 2006) Research indicates that illegal parking, primarily by commercial motor vehicles (CMVs), leads to substantial capacity losses on roadways, particularly during peak times (Han et al., 2005) In response, policymakers are exploring strategies to manage truck and commercial vehicle traffic in central business districts, such as value pricing and off-peak delivery programs While these approaches are gaining traction in the United States, their effectiveness remains uncertain.

This study aims to investigate the effects of the OHD shift program by utilizing popular transportation modeling and simulation software While agencies, consultants, and researchers frequently use these tools for modeling and evaluating passenger travel, there has been significantly less focus on applying them to truck and freight studies.

Selecting the right traffic simulation tool depends on the specific needs and budget of the study The Federal Highway Administration has developed the Traffic Analysis Toolbox, which provides guidelines for choosing appropriate traffic modeling tools (FHWA, 2004) Holguin-Veras et al (2001) explored freight modeling strategies relevant to the New York region, while Boile and Ozbay (2005) summarized various transportation modeling tools and strategies used in New Jersey This article offers a brief overview of freight modeling techniques and their practical applications.

Many states and regions have implemented statewide freight planning models, often utilizing discrete event simulation techniques Freight modeling is recognized as more complex than passenger modeling due to the numerous and dynamic agents influencing freight transport These models can take various forms, including input-output, trip-based, and commodity-based models, with trip-based models resembling traditional transportation planning approaches by focusing on individual trips.

Vehicle travel patterns are analyzed through trip generation rates and their distribution, similar to passenger car models In contrast, commodity-based models focus on the flow of goods in the market, effectively capturing the intricate interactions that characterize freight transportation (Holguin-Veras et al., 2001).

Various agencies have created freight models primarily at the statewide level, with notable examples from states such as Florida, Kentucky, and Texas (Boile et al., 2004; Giuliano et al., 2007) Unlike these states, New Jersey, which encompasses a significant part of New York City's metropolitan area, initially integrated freight operations into its statewide travel demand model instead of developing a standalone freight model Subsequently, New Jersey established a multi-commodity model utilizing GIS tools (Boile et al., 2004) While these freight planning models are beneficial for freight demand management studies, they cannot fully assess the impacts on all traffic types without a comprehensive model that accounts for all vehicle classes.

Travel demand models play a crucial role in transportation planning by predicting travel patterns across cities, states, or regions These models, often developed for metropolitan planning organizations, assess the impact of transportation improvements, forecast future travel, and conduct air-quality analyses The foundation of most travel demand models is the four-step planning process, which includes trip generation, trip distribution, mode choice, and traffic assignment To streamline this process, various commercial software tools have been created, benefiting large networks Notable tools used by government agencies and transportation consultants include TransCAD, CUBE/TP+/TRANPLAN, EMME/2, TRANSIMS, and VISUM.

Modeled transportation networks typically encompass all major roads, highways, and sometimes rail and bus lines Travel demand is influenced by factors such as population, employment, and characteristics of various sub-zones within the area Traffic assignments are conducted for different times of the day to predict future traffic conditions; however, large-scale regional models often utilize static traffic assignment, meaning that assigned volumes are not time-dependent Instead, they aggregate vehicle flows over the entire period, resulting in every vehicle being represented as simultaneously present on each link it utilizes during that timeframe (Boile, 2005).

Regional travel demand models aim to capture all travel activities within a region, tailored to the developers' requirements These models can focus solely on passenger vehicles, freight transportation, or a mix of both Traditionally, freight planning has received significantly less focus compared to passenger car models, resulting in a limited number of developed freight planning models According to Boile and Ozbay (2005), this disparity highlights the need for enhanced freight planning methodologies.

At present, there are no transportation modeling methods that effectively integrate both passenger and freight considerations However, certain transportation planning packages reviewed possess the flexibility to incorporate freight flow data.

Typically, however, for these types of applications, major modifications of the existing models are required and caution should be exercised to develop meaningful models.‖ (Boile, 2005)

7.2.1.3 Passenger and freight planning models

To effectively evaluate the impact of a freight-targeted policy program, it is essential to analyze the effects on all vehicle classes and the overall transportation system While rail transport handles a portion of domestic freight, trucks and delivery vans account for a significant share, utilizing the same highway infrastructure as other vehicles This study specifically examines trucks and delivery vans (CMVs), highlighting that any alterations in their travel patterns will influence the entire highway network and subsequently affect passenger travel Therefore, a comprehensive model that incorporates both passenger and freight travel is necessary to accurately assess traffic impacts across the region.

In recent years, the New York Metropolitan Transportation Council (NYMTC) developed the Freight Planning Model (BPM) in collaboration with neighboring Metropolitan Planning Organizations to enhance regional freight planning for the New York metropolitan area The BPM is a comprehensive travel demand model that encompasses all facets of vehicular travel, including highway freight This integrated approach allows for efficient regional travel analysis by combining passenger car and freight components Moreover, it enables simulations to assess the impacts of various policies, such as toll increases, congestion pricing, and shifts in freight activity based on time of day.

Research Methodology

A methodology has been established to convert the findings of the BMS into scenarios that explore potential shifts in OHD, aimed at analyzing the traffic impacts of the OHD program This includes a summary of key behavioral findings, an outline of the research methodology for traffic models, the construction of scenarios depicting possible demand shifts, and a discussion of model-specific challenges for both employed models.

7.3.1 Summary of relevant behavioral findings

The BMS studies serve as the analytical foundation for this research, focusing on the willingness of food and retail industry receivers and their carriers to engage in Off-Hours Delivery (OHD) Key data includes estimates of truck traffic in New York City, particularly in Manhattan, highlighting the percentage of receivers willing to accept tax incentives to shift deliveries to off-hours (7 PM - 6 AM) This broad-based incentive program gathers insights from businesses in Manhattan's food and retail sectors, with behavioral estimates organized by ZIP codes and simplified into four main community board groupings, as illustrated in Figure 85.

Figure 85: Community Board Districts in Manhattan

To effectively simulate traffic, the research team must determine the proportion of commercial traffic in the study area that pertains to food and retail deliveries, as impacts on these industries are the focus of the study Different industries exhibit varying responses to incentives, affecting their share of overall truck traffic Consequently, only food and retail-related commercial traffic is anticipated to alter its behavior, prompting the need to estimate the likelihood that a commercial motor vehicle (CMV) in the traffic stream is engaged in food or retail deliveries This estimation utilizes trip generation models developed by the BMS research team The analysis includes a breakdown of deliveries by each Manhattan community board district in Table 16, while Table 17 further categorizes these deliveries by industry, specifically food and retail, as a percentage of total deliveries.

Table 16: Delivery Percentages for Community Board District Groupings

Total Food Deliveries for Manhattan 46,113 100.00%

Table 17: Proportion of Truck Traffic by Industry

Based on the data from Table 17, we can calculate the total percentage of Commercial Motor Vehicle (CMV) traffic that would respond to the proposed incentive program, assuming all CMV trips to Manhattan are deliveries Behavioral studies provide estimates on the percentage of truck traffic likely to transition from daytime to overnight operations To effectively analyze this shift, a model is developed to incorporate the Overnight Delivery (OHD) program into the selected traffic models, integrating insights from the Behavioral Modeling Study (BMS).

7.3.2 Commercial vehicle (CMV) demand shift model

The BPM and its sub-simulation model provide certain benefits compared to traditional four-step demand forecasting models, particularly in their iterative planning process and micro-simulation features However, they fall short of fully meeting the requirements of this study, as they lack the capability to automatically re-assign and redistribute traffic in response to predicted changes.

The methodology must be adjusted to address the complexities of the models in BPM, where the traffic assignment module operates autonomously In this system, vehicles independently select optimal routes from their origin to destination (Parsons Brinckerhoff, 2005) Consequently, alterations in truck behavior and routing are reflected by adjusting the volume of commercial vehicle trips between locations.

In both models, commercial vehicle trips between origin and destination zones are documented in an OD matrix, where each cell represents the number of trips per time period After modifying the existing commercial vehicle OD matrices, the traffic assignment module of the BPM and the mesoscopic simulation network will be re-executed This allows for a comparison between the results of the traffic assignment based on the adjusted OD demands and those from the base assignment Changes to the CMV OD matrices are implemented using shift factors derived from behavioral data.

Manhattan, the commercial hub of New York City, boasts the highest population density and traffic levels in the area, making it the primary focus of various traffic control initiatives, including congestion pricing The borough is divided into districts based on geography and commercial density, and due to its island status with limited access points, Manhattan is particularly straightforward to analyze from a transportation modeling perspective.

The detail and size of transportation analysis zones in traffic simulation models are influenced by two main factors: the zone system of the model itself and the system utilized by behavior modules The Behavioral Pattern Model (BPM) utilizes a zone system based on census tracts, resulting in 3,586 transportation analysis zones across the New York region, with 2,374 zones in New York City and 318 in Manhattan (Parsons Brinckerhoff, 2005) Delivery data for unique ZIP codes in Manhattan is available by market segments, allowing these ZIP code areas to be organized by community boards into four general zones The zones from each traffic model can be similarly categorized to align with these community board groupings.

Shift factors derived from behavioral data are utilized to analyze Commercial Motor Vehicle (CMV) origin-destination (OD) demands from various zones outside Manhattan to destinations within the borough Additional simulations focus specifically on OD demands targeting Downtown and Midtown Manhattan, collectively referred to as "Lower Manhattan." This region encompasses community board groupings 1-6 and includes Manhattan's primary business districts, which house the majority of the city's commercial establishments Notably, over 75% of all deliveries in Manhattan occur within these key areas.

The shift factors, α J , developed from the behavioral data are used to factor the commercial vehicle origin-destination demand, x ij , as shown in Equation 19:

( ) ( ) p p ij new ij old j x  x   (19) where x ij = CMV trip demand between origin ‘i' and destination ‘j’ α J = shift factor for trips with destination in zone J

2 for Midday Period p 3 for PM Peak Period

The BMS results are applicable to all daytime hours, as they are not restricted to specific time periods Consequently, the same α J factor is utilized for the demands across the three daytime segments.

This study aims to model and evaluate the effects of shifting freight traffic from daytime to off-peak hours It is assumed that all freight traffic reduced from the three daytime periods will be redirected to the overnight off-hour period Consequently, the total daily demand for commercial vehicles between any origin-destination pair (X ij) remains unchanged throughout the 24-hour period, both in the existing and shifted scenarios, irrespective of the values of α J.

  is constant for all ij pairs (20)

In period 4 (overnight), no α J factor is utilized; instead, the overnight demand is determined by summing the existing demand with the shifted demands from the other three time periods Consequently, for each OD pair, the new overnight off-hour demand is computed using Equation 21.

( ) ( ) J ( ) ( ) ( ) ij new ij old ij old ij old ij old x  x   x  x  x (21)

A destination zone is the endpoint of trips within a transportation model, where trips are organized in an origin-destination (OD) matrix before being allocated to the network To apply a shift factor to a specific group of zones, all OD pairs that have a destination within the selected group of J zones will be assigned this factor.

Freight traffic from various origins heading to Manhattan will experience a shift to off-peak hours This adjustment is implemented through a MATLAB script that requires modifications to reflect the updated zones in the model associated with these trip shifts.

Network Calibration

This section outlines the initial investigation and calibration of two traffic models utilized in this study The calibration process finalized all input parameters for the simulation models, which are essential for simulating the proposed programs The first part examines the behavior of the BPM and implements the research methodology, while the second part focuses on calibrating the commercial vehicle origin-destination matrices The third part involves extracting and calibrating the Manhattan sub-network for use in the mesoscopic simulation tool, TransModeler Following initial testing to validate the models' results, a comprehensive calibration procedure was conducted using the latest available data to ensure accuracy Completing these tasks paves the way for the successful implementation of the research methodology and the achievement of the study's objectives.

Initial testing of the BPM demonstrated strong performance at a macroscopic level, leading the research team to recognize it as a credible model due to its extensive application as an evaluation tool in numerous studies within the New York metropolitan area Ongoing development and calibration efforts by NYMTC continue to enhance its effectiveness.

The research team dedicated significant effort to developing and testing a reliable version of the model, as outlined by Parsons Brinckerhoff (2005) Initially, the behavior model was incomplete, leading to the use of arbitrary shift values and manipulated OD demands for the traffic simulation The results from these tests were then compared to existing baseline scenarios to assess the model's responsiveness to changes and establish a foundation for future testing and modifications.

7.4.1.1 Predicted link flow and travel time changes

The full BPM model produced unexpected results, contradicting the assumption that reduced commercial vehicle (CMV) demand would lead to decreased congestion Instead, the model indicated that lower CMV traffic increased auto and other vehicle trip generation rates, exacerbating congestion on a network already nearing saturation This surge in auto trips compensated for the decline in CMV traffic, resulting in higher daytime travel times and congestion levels Consequently, the research team was prompted to undertake a more in-depth analysis of the BPM's trip generation and assignment processes.

Before comparing assignment results, trip generation analysis reveals the impact of changes in commercial vehicle (CMV) demand on non-commercial vehicle (auto) traffic The iterative nature of the BPM model shows that auto trip generation rates increase during AM Peak, Midday, and PM Peak periods as CMV traffic decreases Interestingly, this trend also extends to the Night period, contrary to expectations that shifting CMVs would reduce auto trips This anomaly arises from BPM's pre-assignment processor, which uniformly distributes generated trips throughout the day, irrespective of actual traffic conditions Consequently, while decreased truck activity leads to more auto trips, the model's consistent distribution pattern highlights a limitation in its applicability to this study, prompting a need for methodological revisions.

Figure 89: Change in Auto Trip Generation due to CMV Shift

7.4.1.3 Adapting research methodology to BPM

The New York Metropolitan Transportation Council's Best Practice Model, as detailed in the Final Report Documentation by Parsons Brinckerhoff (2005), outlines a trip generation component that aggregates "journeys" based on various factors such as destinations, stops, and transportation modes throughout the day Following this, a time-of-day processor transforms these journeys into individual "trips" between zones, enabling their representation as an Origin-Destination (OD) matrix in TransCAD This conversion relies on cumulative distributions for arrival times and activity durations associated with each journey purpose.

The aggregated trips for each journey type, with their assigned arrival times, are split into trips within

The analysis utilizes a 76 x 48 matrix to weight 48-hour time periods, categorizing trips into four model time periods: AM Peak, Midday, PM Peak, and Night, while distinguishing between six highway vehicle classes and four transit classes Each trip is confined to its respective time period without overlap, implying that identical Time of Day (ToD) factors are applied consistently, irrespective of variations in trip generation and mode split Consequently, any fluctuations in trip generation result in proportional adjustments across all time periods based on the established factors.

The model indicated an overall increase in daily trips due to reduced travel times in three out of four time periods and fewer truck trips, leading to heightened activity Consequently, consistent time-of-day shifts applied to trip distribution resulted in uniform percentage increases in demand across all four time periods The developers of BPM highlighted these findings in their final report.

Incorporating more flexible timing considerations into travel modeling systems is essential for accurately reflecting individual travel patterns, including journey sequencing and scheduling This enhancement would also increase the system's responsiveness to policy measures designed to alleviate congestion.

To address the shortcomings of the BPM time-of-day splits in accurately modeling truck time-of-day shifts, several strategies were evaluated to enhance the existing framework while leveraging the strengths of BPM.

7.4.1.3.1 Assume non-commercial trips to be constant

To enhance model efficiency, a proposed solution is to prevent the re-execution of trip generation and mode splits after a certain percentage of commercial vehicle traffic has been shifted Instead, shifts in commercial motor vehicle (CMV) traffic would occur externally to the model, with only the highway assignment module being recalibrated for each scenario This approach allows for variations in assignments while disregarding the trip generation, distribution, and mode split modules.

OD matrices will be manually adjusted prior to assignment, providing realistic outcomes for short-term analyses as immediate shifts in truck traffic do not instantly alter travel patterns or trip generation However, this approach limits the model's full execution, relying solely on TransCAD's highway assignment features While it effectively estimates immediate traffic impacts from commercial vehicle shifts, it does not account for long-term trends or changes in trip generation resulting from altered network conditions.

7.4.1.3.2 Manipulate time-of-day factors

The time-of-day (ToD) factors in Business Process Management (BPM) are organized within a predefined matrix that can be manually adjusted for analysis By increasing ToD factors for nighttime freight traffic and proportionally decreasing daytime values, the overall traffic distribution throughout the day can remain unchanged This approach eliminates the need for manual adjustments between daytime and nighttime origin-destination matrices, ensuring consistent trip distribution across all vehicle classes However, challenges arise in determining the precise adjustments to the ToD factors, as there is limited documentation on their initial selection, complicating the manipulation process.

7.4.1.3.3 Manually redistribute newly generated night auto trips to other periods

One effective strategy for managing trip generation involves maintaining the model's current processing procedures while manually redistributing newly generated night auto trips to existing daytime periods This can be achieved by halting the BPM model run between the time-of-day (ToD) processor and highway assignment, allowing for the subtraction of the base level origin-destination (OD) matrix for the night period from the newly generated OD matrix The resulting differences for each OD pair are proportionally added to the newly generated AM Peak, Midday, and PM Peak matrices, based on ToD splits and the daily total OD demand percentage for each time period This method effectively accommodates increased truck demand while preserving existing night auto levels; however, it presents challenges in complexity and clarity regarding the distribution of demand across different time periods.

Due to the complexity and lack of clear methodologies for model manipulation, the research team opted to maintain constant auto trip generation for the BPM run results in the short term This approach not only streamlines the modeling process but also significantly reduces the time required for traffic assignment from several days to just a few hours, allowing for the analysis of multiple scenarios While the advantages of TransCAD's highway assignment procedures are utilized, it is important to note that only the assignment portion of the BPM is applied.

7.4.2 Truck origin-destination matrix calibration

Simulation Results for Broad-Based Incentive Policy

A methodology was established to create an off-hour delivery program tailored for food and retail businesses in Manhattan, utilizing a BPM and a mesoscopic simulation sub-network This involved reallocating truck and commercial van trips destined for Manhattan from three daytime periods (AM Peak, Midday, PM Peak) between 6 AM and 7 PM to the nighttime period from 7 PM to 6 AM The study modeled twelve scenarios of a broad-based incentive program for the food and retail sectors, with three scenarios specifically simulated in the Manhattan sub-model The results presented are aggregated from the BPM highway assignment and mesoscopic simulation analyses.

7.5.1 Commercial Vehicle shift model results

The CMV shift model was utilized to transfer the OD demands of commercial vehicles, including trucks and vans, from daytime periods to the overnight period, specifically targeting trips destined for Manhattan This model accounts for both direct trips originating in Manhattan and chained deliveries from outside the city, which involve multiple stops within Manhattan A detailed analysis reveals that the majority of commercial trips to Manhattan come from the other four boroughs of New York City, with a significant number also originating from New Jersey and points west Trips from north of the city represent just under 10% of the total, while the fewest originate from Long Island to the east.

Figure 103: Origin of CMV Trips Destined for Manhattan

The shift factors outlined in Table 18 were uniformly applied to all Commercial Motor Vehicle (CMV) trips across the New York region, irrespective of the origin zone or time of day (6 AM - 7 PM) The total number of shifted trips and their percentage relative to all CMV trips in the region are detailed in Table 30 for the BPM all-Manhattan scenarios Additionally, the correlation between the scenario and the number of trips shifted is illustrated in Figure 86.

Table 30: CMV Trips Shifted – All-Manhattan Scenarios Sce nario CMV Trips Shifte d % of all CMV Trips

The resultant matrices from the shift model were input to BPM highway assignment and the mesoscopic simulation The network results are shown in the following sub-sections

The macroscopic network model's (BPM) assignment output provides detailed information on over 55,000 links in the highway network, including vehicle flows by class, travel times, and average speeds Two key parameters derived from this data are Vehicle Miles Traveled (VMT) and Vehicle Hours Traveled (VHT) VMT represents the total distance traveled by all vehicles in the region on a typical day, while VHT serves as an effective measure of travel times and congestion levels Although variations in VMT do not directly indicate changes in congestion, analyzing VHT allows for a clearer understanding of traffic conditions.

190 example, vehicles may take longer paths to avoid congested links, and in turn reduce their overall travel time, thus saving time and reducing VHT while increasing VMT

The analysis presents the net differences in output parameters between the calibrated 2007 base model and the shift scenario model, along with the percentage changes observed Specifically, Figures 104 and 105 illustrate the variations in Vehicle Miles Traveled (VMT) resulting from the assignment of shifting scenarios, focusing on cases where all Manhattan-bound demands were shifted, as well as those where only Lower Manhattan (Midtown and Downtown) demands were affected Additionally, Figures 106 and 107 depict the changes in Vehicle Hours Traveled (VHT) for these two scenarios.

Figure 104: Change in VMT – All Manhattan Destinations Shifted

ECONOMIC IMPACTS