Transportation Systems Planning Methods and Applications 16 Transportation engineering and transportation planning are two sides of the same coin aiming at the design of an efficient infrastructure and service to meet the growing needs for accessibility and mobility. Many well-designed transport systems that meet these needs are based on a solid understanding of human behavior. Since transportation systems are the backbone connecting the vital parts of a city, in-depth understanding of human nature is essential to the planning, design, and operational analysis of transportation systems. With contributions by transportation experts from around the world, Transportation Systems Planning: Methods and Applications compiles engineering data and methods for solving problems in the planning, design, construction, and operation of various transportation modes into one source. It is the first methodological transportation planning reference that illustrates analytical simulation methods that depict human behavior in a realistic way, and many of its chapters emphasize newly developed and previously unpublished simulation methods. The handbook demonstrates how urban and regional planning, geography, demography, economics, sociology, ecology, psychology, business, operations management, and engineering come together to help us plan for better futures that are human-centered.
16 Centre SIM: First-Generation Model Design, Pragmatic Implementation, and Scenarios JoNette Kuhnau Pennsylvania State University Konstadinos G Goulias Pennsylvania State University 16.1 16.2 16.3 16.4 CONTENTS Introduction Study Area, Issues, and Background Centre SIM Background Centre SIM Scenario Testing Description of Scenarios • Network Changes • Staggered Work Hours Policy Scenario Evaluation 16.5 Summary and Next Steps References 16.1 Introduction Regional transport system simulation is most often based on a computerized model system that contains the resident population’s social, demographic, economic, and location characteristics The system also contains statistical models of travel behavior that use as input the population characteristics to predict the number of trips people make and the places among which these trips are made, the means used to travel among these places, and the routes chosen to get from one place to another One system designed to this in sequence is called the four-step model system, and it has been under continuous scrutiny for the past 30 years (e.g., see the resource papers from the 1972 conference on Urban Travel Demand Forecasting, published by the Highway Research Board, now named the Transportation Research Board, in Special Report 143) In order to depict and compute the routes chosen, the roadway and public transportation network with its characteristics (e.g., roadways and terminals) are also needed In addition, many model systems also incorporate urban and rural spatial attributes, representing locations where persons pursue their everyday activities (e.g., work, education, eating meals, shopping) All this information is arranged in an electronic map to represent the spatial separation of places and the spatial organization of the transportation network connecting these places © 2003 CRC Press LLC For modeling purposes, a study area is usually identified based on administrative and institutional criteria (e.g., a county, a city), but very often functionality is also taken into account (e.g., the region that contains Philadelphia, Pennsylvania, contains geographic portions in the state of New Jersey) This electronic map with embedded databases and behavioral equations constitutes software in which scenarios of change can be imposed (e.g., building a new road and therefore modifying the road network, or building new residences and therefore changing the spatial distribution of activity locations) and their effects computed by the embedded models Outputs of the system include the number of vehicles per hour (volumes) traveling on each element of the network (links and nodes) and a variety of performance measures, such as vehicle kilometers of travel and travel speeds In addition, postprocessing of this information can be used to derive air pollutant emissions Forecasting is performed by estimating expected changes in the use of land, demographics, and network characteristics, incorporating these changes into the digital map, and running the entire software application to predict future traffic volumes Model building like the above for long-range transportation planning, has been very important, but it is becoming increasingly more important for cities and metropolitan planning organizations (MPOs) for short-range operational analysis For example, cities also need tools that help them to maintain building facility and asset inventories and to study their policies a year ahead (e.g., parking changes and signal coordination), and provide data to study operational improvements in public transportation To this, the state of the practice in planning for transportation engineers is still the four-step urban transportation planning system (UTPS) model consisting of trip generation, trip distribution, mode choice, and traffic assignment These sequential models require large quantities of data, significant time investment, and considerable computer resources, even for small urban areas Yet, they are loaded with well-known deficiencies such as lack of behavioral consideration, untested assumptions about uniform trip-making behavior within artificial zones, lack of time-of-day considerations, and so forth Because these models have little or no behavioral basis, they cannot be used to evaluate the effects of transportation demand management strategies and other programs and small improvements in the transportation network that may have effects that are smaller than the model’s error Further, the models have little or no temporal resolution, making time-dependent issues, such as emissions estimates, impossible This capability has become critical for regions with severe transportation problems, such as areas classified as air quality nonattainment areas (see Chapter 13 on emissions issues and Chapter on behavioral issues) Even as researchers develop improved travel demand methods, in practice there is not sufficient experience or resources for new data collection, analysis, and model development This leaves academia and practicing transportation planners with an ever-increasing gap between theory and practice Person- and household-based activity surveys are becoming more widely used, but small metropolitan planning organizations are not prepared to discard their current models and start over with activitybased models Even as activity-based models become more accepted in practice, the resources for data collection, model development, and calibration are not available for the striking majority of small metropolitan areas To assist planners in moving from the traditional models to behavior-based methods, transitional approaches are needed that can incorporate activity and behavioral data without abandoning the models that are well understood and available This is particularly critical for urban areas that not have the resources for data collection and for which a complete transition to a new model is not yet possible In this chapter the first version of a model system (Centre SIM) incorporating activity information in a traditional UTPS model is presented Centre SIM improves the accuracy and predictive capabilities of the traditionally applied four-step model This is based on a simple and practical methodology that effectively incorporates person-based activity behavior within the structure of a four-step model The remainder of the chapter is organized as follows First an overview of the study area, issues, and background is provided This is followed by a summary of the Centre SIM model effort and a set of scenarios studied with the model The chapter concludes with a brief summary and a discussion on the next steps © 2003 CRC Press LLC 16.2 Study Area, Issues, and Background The study area for the Centre SIM model is Centre County, Pennsylvania, a region of approximately 136,000 persons Centre County includes the main campus of the Pennsylvania State University (Penn State), with over 50,000 students, faculty, and staff The presence of Penn State dominates the time-ofday travel patterns of the region because a large number of residents (faculty and students) have flexible activity and travel schedules, creating an observed pattern of peak traffic flow in the evening rather than in the morning It is therefore important to consider the special characteristics of the Penn State community when simulating the Centre County population The University Park Campus Master Plan was approved by the Pennsylvania State University in 1999 The plan outlines the development and redevelopment of the University Park campus (in Centre County) over the next 25 years The master plan contains provisions for substantial construction of new classroom and laboratory facilities, but it is fundamentally based on the concept of an accessible campus with a student-oriented campus core The Penn State Master Plan Transportation Committee (MPTC) was subsequently formed to study ways in which to achieve the environment- and pedestrian-friendly campus described in the master plan One of the visions is exploring the effects of projects aimed at making the core of the campus pedestrian friendly (e.g., see the intermodal transportation concept at http:// www.opp.psu.edu/divisions/cpd/trans/ITChome.htm (accessed May 2002)) In addition, university growth requires a rearrangement of the parking structures The effects of moving parking lots and construction of additional parking need to be computed and added to the traffic flow impacts In parallel, Centre County is experiencing major additions to its roadway network and significant growth and land use changes Therefore, the objective of this application example is to study and assess the effects of the campus projects, while at the same time accounting for all other changes taking place in the county Figure 16.1 shows the study area’s location in Pennsylvania, its jurisdictional composition (the southernmost municipalities are the most populated), and identifies one major roadway construction project that is under way with the objective of creating a four-lane controlled access (freeway) connection between the region and Interstate 80 as part of Interstate 99 16.3 Centre SIM Background In the past, a transportation demand model for Centre County, within an initiative called Access Management Impact Simulation (AMIS), was developed to aid in operational studies of the region, such as impact fee assessment, signal coordination, and traffic calming (Chung, 1997) At that time, the model was intended as a more detailed representation of development to help public agencies assign fees based on traffic contribution by each development (Goulias and Marker, 1998) and to study traffic engineering highway improvements and demonstrate their impacts by exploiting the capabilities of Geographic Information Systems (GIS) (Chung and Goulias, 1996) In 2000, this model was further updated, improved, and calibrated with the specific objective to support the Master Plan Transportation Committee in its deliberations This model encompasses all of Centre County with 1067 traffic analysis zones (TAZs), 3458 roadway links, and 3073 nodes The digital map is divided into TAZs in its more densely populated region that are based on census blocks (equivalent to city blocks), while other TAZs are based on census block groups (in rural areas with low residential densities) In AMIS the amount of travel each household produces is estimated at the household level using other surveys (Chung, 1997) Cluster analysis was used to create groups of households with similar trip-making patterns in a day The day was divided into ten time periods, and the number of trips for each household type in each period was estimated The evolution of households was also simulated in a sociodemographic prediction model from 1980 to 1990 (for validation purposes), and then simulated again to 1997 The number of trips attracted by each business site in the county was calculated at the individual business level based on 1997 data The data included number of employees for each business and type of business, with trip rate equations extracted from the widely used Institute of Transportation Engineers (ITE) Trip Generation Manual The time-of-day profiles © 2003 CRC Press LLC FIGURE 16.1 Centre County in its geographical context for customer arrivals were derived from a telephone survey of the businesses in Centre County In this way, the amount of traveling (trip generation rates) is estimated at the household and business levels, which is better than zonal trip generation because it does not require the assumption of uniform characteristics throughout the TAZ In a parallel study, the Penn State campus was used to demonstrate the feasibility of an activity-based evacuation management model system, and for this reason, more detailed data were used to study time allocation on campus (Alam, 1998) To this end, an activity survey was conducted in October and November 1996 and was used to determine the number of Penn State faculty, staff, and students traveling during different periods in the day An example of this time allocation in a time slice of the 24 periods in a day and for each activity type is provided in Figure 16.2 These activity data were used in conjunction with other parking data to determine the number of vehicle trips originating and ending in the zones on campus Alam’s study made the key connection between activity modeling and the four-step framework using a building presence model (this is the key idea used in developing Centre SIM to a zone presence model covering the entire county) Activity patterns for the different population segments provide probabilities for activity participation at time points (buildings on campus) throughout the day By combining this activity participation information (in Figure 16.2) with land use data (e.g., the detailed information about the use of buildings on campus), the geographic distribution of persons in the study area is produced Figure 16.3 provides an example of time-of-day allocation to activity locations in Alam’s model Each cubical solid shape in the map of Figure 16.3 is the amount of persons in that specific building In this way, Alam built 24 maps (1 for each hour in a day) that depict campus life (Alam, 1998) As part of the Penn State Master Plan project, the original AMIS model was updated by adding network (roadway) and development changes that occurred between 1997 and 2000 Development changes © 2003 CRC Press LLC 100 90 Time Spent on Activities (%) 80 70 60 50 40 30 20 10 L 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.3 0.0 0.0 0.0 0.0 0.6 0.6 0.0 K 0.0 1.0 1.2 0.3 0.0 0.0 0.0 5.2 8.4 10.0 6.4 8.3 11.3 9.0 10.7 11.2 9.6 10.4 9.9 8.4 6.8 8.0 3.5 2.8 J 8.4 2.7 0.0 0.0 0.0 0.0 0.0 0.5 1.7 0.6 1.3 1.5 2.2 1.0 0.6 0.1 1.3 4.6 7.3 15.1 12.0 10.3 11.1 12.9 I 7.7 5.6 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.4 2.6 1.8 1.1 2.0 2.6 2.8 2.1 3.6 5.7 12.4 18.9 23.7 22.2 18.2 H 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.6 1.2 1.3 0.0 0.0 1.2 0.2 0.0 0.0 0.0 0.0 0.5 1.9 1.3 0.0 0.0 G 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.6 0.0 0.2 0.0 1.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 F 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.4 0.0 1.4 1.3 0.0 0.2 0.2 1.3 4.0 5.4 2.4 2.4 0.0 1.0 0.0 E 0.2 0.0 0.0 0.0 0.0 0.0 0.0 0.1 1.0 0.5 1.3 1.2 3.0 2.9 4.3 4.1 4.2 3.7 5.2 7.9 3.5 4.2 4.3 1.3 D 20.5 10.9 7.1 2.2 1.3 1.3 0.0 3.7 20.2 36.3 51.0 52.1 47.3 44.0 45.1 45.5 45.1 31.8 33.4 32.9 35.2 31.2 31.9 26.6 C 1.3 1.0 0.0 0.0 0.0 0.0 0.0 1.0 5.7 15.6 16.6 20.8 17.5 24.9 25.6 27.7 24.0 17.6 9.7 7.5 6.5 6.8 6.6 0.0 B 0.0 1.6 0.0 0.0 0.0 A 61.9 77.2 90.4 96.2 97.4 97.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.6 0.6 5.8 6.0 9.6 6.2 7.9 14.4 11.6 6.9 3.4 2.5 16.9 16.5 8.3 8.2 4.7 1.7 0.0 98.1 82.5 54.7 24.8 10.8 5.1 2.5 3.3 3.5 3.3 5.6 5.9 10.0 4.7 7.1 8.9 18.6 35.7 Time Segment (Hour) FIGURE 16.2 Students’ time allocation in the Penn State survey FIGURE 16.3 The Alam spatial and temporal time allocation model (12:00 noon to 1:00 P.M.) included construction of commercial areas and major residential areas The calibration of one model version was performed by assigning traffic to the network and comparing the resulting volumes to peak period traffic counts For each traffic assignment run, differences between volumes predicted by the model and actual volumes were noted Adjustments were subsequently made to improve the model system and minimize these differences This was deemed insufficient for studies that require detailed information about time-of-day travel and network volumes for minor roadways Different options for improvements were explored, targeting incorporation of Alam’s approach in a county-wide frame and increasing the level of detail usually found in four-step models Kuhnau (2001) developed this new framework for simulating the entire county for each of the 24 daily segments by adding some key information and simplifying the Alam activity model This model system is named Centre County Simulation (Centre SIM) Figure 16.4 provides a flowchart of all the information sources and modeling steps that are a combination of the Alam procedures to assign persons at geographic © 2003 CRC Press LLC PSU Activity Survey PSTP Wave Activity Patterns by Time of Day (TOD): Students, Faculty, Staff, Unemployed Centre County Business List Address Matching Convert SIC Code to Activity Types Probability of Travel (by TOD and Activity) Activity Capacity of Businesses Aggregate Business Capacities to TAZ PSTP Wave TAZ Presence (by Activity, Category and TOD) Mode Split (by Category and TOD) Origins and Destinations for TAZs by TOD AADT & NCHRP 187 for Gate TAZs Peak Network Midday Network Trip Distribution TAZ Capacity for Each Activity GIS Network of Centre County, PA Off-Peak Network Traffic Assignment Traffic Volumes by TOD and Direction (passenger cars only) Link Travel Times by TOD Traffic Counts Calibrate/ Validate FIGURE 16.4 Centre SIM methodology flowchart locations pursuing activities, and the four-step model system Below we identify a few key aspects of this model effort The primary data elements used in this Centre SIM are: • • • • • • • Penn State activity survey Puget Sound Transportation Panel (PSTP) wave (data from year 1997) 2000 U.S Census population statistics 2000 business list for Centre County Hourly and daily traffic counts GIS network of Centre County Link travel times for three traffic conditions (peak, midday, off-peak) An activity survey of a representative sample of the Centre County population, not only the Penn State campus, would have been more appropriate for this application However, this was not available during model design and testing, so a much smaller survey designed by Alam, together with data from other regions, was used This has some important implications, as discussed in the summary and conclusions For simulation of activity patterns, the population of Centre County greater than 18 years of age was divided into the following six categories: • Penn State students Penn State faculty â 2003 CRC Press LLC • • • • Penn State staff Unemployed persons Professionals Workers (Non-Penn State staff) Each of these categories was assumed to have different activity and travel patterns affecting the overall travel demand observed It is important to note that persons under 18 years of age were not included in this model due to lack of information about activity patterns and employment rates This may not be a serious deficiency because we focus on passenger travel by car to demonstrate feasibility and identify the next steps From 2000 U.S Census data, only about 2500 persons are between 16 and 18 years old, less then 2% of the total population of the county As mentioned above, the person-based activity patterns used in this model were derived from a 2-day activity diary of 102 Penn State students, faculty, and staff These activity patterns were modified slightly for persons employed outside the Penn State campus This was done because of the unique characteristics of the location and constraints of the Penn State campus in relation to State College Figure 16.5 shows the time allocation of students by time of day Kuhnau (2001) developed an algorithm to convert these time allocations by each population segment to time allocations for the entire population Then these population time allocation shares by activity type are converted to time allocation at destinations To this, another algorithm is needed that is in essence a distribution model based on capacity and type of activity that each business location allows Travel is then computed using a series of steps that are similar to the fourstep procedures A key difference here, however, is that the spatial and temporal distribution of activity types is used to derive the propensity to travel and modal share by each of the population segments above Figure 16.6 is one of the outputs obtained from the model, and it contains both the spatial distribution of persons in activities and the traffic volumes predicted Model validation and calibration were also needed for this type of model, in a way similar to that of other four-step applications As Kuhnau (2001) shows, however, a model that departs from activity participation and contains within it the time-of-day variation in travel is closer to observed counts and requires less drastic calibration steps FIGURE 16.5 Time use pattern for students (From Kahnau, J.L., Master’s thesis, Pennsylvania State University, University Park, 2001 With permission.) © 2003 CRC Press LLC FIGURE 16.6 Centre SIM activity participation (zonal presence) and travel in the period 5:00 P.M to 6:00 P.M 16.4 Centre SIM Scenario Testing This section describes how the Centre SIM model was used to assess proposed transportation projects and plans on the Penn State University Park campus The scenarios include network (roadway) changes, as well as a travel demand management (TDM) proposal to spread the Penn State peak hour travel demand The results of these scenarios are discussed from both the modeling and planning perspectives The objectives of scenario testing with the Centre SIM model were to demonstrate the capabilities of the model, evaluate alternative plans proposed by Penn State University, and analyze possible solutions to the localized congestion problems around the Penn State campus First, a base scenario of existing conditions was simulated to provide a benchmark against which the simulated alternatives were evaluated The changes proposed in each scenario were then created, simulated, analyzed, and compared to determine the effects of different proposals on the transportation network on the Penn State campus and in the surrounding State College area The versatility of the model was demonstrated by its ability to model a wide variety of proposals and produce interpretable results necessary to make recommendations about the feasibility of each project 16.4.1 Description of Scenarios As mentioned in the first section, many projects are envisioned in the Penn State Master Plan Specifically, several proposed parking and roadway changes required simulation to determine their potential impacts on traffic volumes and vehicle circulation patterns on campus The simulations were designed to provide a quantitative basis for decision making about the feasibility of further planning for each alternative Road closures and construction obviously will have an effect on traffic flows on campus In addition, the construction of new buildings brings increased demand for parking facilities, while at the same time it removes surface parking capacity The three proposed roadway changes evaluated in this research are described below and shown in Figure 16.7: © 2003 CRC Press LLC FIGURE 16.7 Penn State roadway scenarios Network scenario 1: Closure of Shortlidge Road Involves closing one block of the only direct north–south route through the core campus, connecting downtown State College to Park Avenue (a major east–west route bordering the Penn State campus) Network scenario 2: Extension of Bigler Road Extends an existing route to College Avenue, creating a new direct north–south link through campus from Park Avenue to downtown State College Network scenario 3: Conversion of Curtin Road to transit only Includes prohibiting all but transit vehicles from a section of roadway heavily used by transit The route connects the easternmost parts of the campus (including major residence and parking facilities) with the westernmost areas (consisting of classrooms, laboratories, and offices) Another scenario tested was defined to examine the model capabilities when the time-of-day aspects of the model are used The proposed TDM policy scenario involved staggering the work starting and ending times for Penn State staff in an attempt to mitigate the severe demands placed on portions of the transportation network at 8:00 A.M and 5:00 P.M For this scenario, it was assumed that one third of the staff would start at the hours of 7:00, 8:00, and 9:00 A.M Consequently, the corresponding work end times would be approximately 4:00, 5:00, and 6:00 P.M No assumptions were made about changes in travel during the period around 12:00 P.M (lunch) because it was unclear how staff would change their midday behavior due to different work starting times In addition, no changes were made to the faculty or student activity patterns because they not have fixed schedules dictated by a single employer Thus, the hours simulated for the staggered hours policy were 6:00, 7:00, and 8:00 A.M for the work arrival patterns because, for example, travel must occur in the 6:00 to 7:00 A.M segment for a staff member to be engaged in work at 7:00 A.M The 4:00, 5:00, and 6:00 P.M hours were simulated for the work departure pattern since if work ends at 4:00 P.M., a trip to the next activity will occur in the 4:00 to 5:00 P.M time period (the same time segment as the end of the workday) 16.4.2 Network Changes Modifications to the transportation network were easily modeled in the GIS software TransCAD, using its algorithms to recalculate the shortest path between an origin–destination pair and reassign the trip route First, three different networks were created to test each of the scenarios individually before examining the effects of combinations of alternatives © 2003 CRC Press LLC FIGURE 16.8 Sample output from Shortlidge Road scenario For network scenario 1, the roadway closure was simulated by removing the appropriate section of Shortlidge Road from the link layer in the GIS map of the model In scenarios to that involved only network changes, it was assumed that the activity patterns, zone presence, travel patterns, and mode split would remain constant Employing these assumptions, the base origin–destination patterns (i.e., the origin–destination matrix) can be used for simulation of all three roadway scenarios This assumption was reasonable because, for example, a road closure of one block changes the route choices available for a trip and may slightly change the travel times between some origin–destination pairs, but would be unlikely to influence the decisions whether to make a trip, what mode to use, and where to go (trip generation, mode split, and trip distribution) A new link was added to connect the existing section of Bigler Road to College Avenue, simulating scenario Characteristics such as travel time, number and capacity of lanes, and directions of flow were assigned to the link to make it usable for traffic in the GIS network Next, the link designated as transit only was removed to simulate scenario Personal vehicles only, not transit vehicles, were simulated in the Centre SIM model; therefore, creating the transit-only section in essence removes it from the nontransit (personal vehicle) network A new network file was created for each of these scenarios for use in the reassignment of traffic flows on the modified transportation system After completing the network changes described, the origin–destination matrix resulting from trip distribution in the base model was assigned to each of the three new networks TransCAD then produces the typical traffic assignment graphic and database output for each scenario An example of the traffic assignment output from the Shortlidge Road scenario (scenario 1) is shown in Figure 16.8 for the 5:00 to 6:00 P.M time segment Although other time segments were simulated, the P.M peak in State College exhibits the most significant impacts because the overall travel demand is highest during this time The result in the P.M peak from the simulation of the Shortlidge Road scenario, shown here, is thus the worstcase outcome of the proposed alternative In this form, however, impacts of the scenario can only be seen through manual link-by-link comparisons between the base and scenario models To create a clearer demonstration of the scenario impacts, Figure 16.9 shows the differences between baseline and the Shortlidge scenario, again for the 5:00 to 6:00 P.M segment (the peak hour) The positive values indicate increases in traffic volumes due to the road closure, while the negative values represent decreases in vehicle volumes The major volume decrease in traffic all along Shortlidge Road was expected because it no longer provides a direct route through campus In addition, one parking deck structure lies on each side of the closure, and as a result, these facilities © 2003 CRC Press LLC FIGURE 16.9 Sample output from scenario impact evaluation can only be entered and exited from one direction if the scenario is implemented This type of comparative analysis was performed for the other two individual alternatives (Kuhnau, 2001) 16.4.3 Staggered Work Hours Policy Scenario Evaluation The staggered work hours scenario, unlike the other three alternatives described, involved behavioral (activity pattern) changes as opposed to physical (network) changes Thus, the same network used for the base condition was also used to evaluate this policy scenario To simulate the changes in arrival and departure times to and from work, a modified activity distribution was created Whereas formerly the distribution of Penn State staff arrivals at work had a concentration of 50% arriving in the 7:00 to 8:00 A.M time segment, the new activity distribution was designed with 30% of the staff category engaged in work at 7:00 A.M., 60% at 8:00 A.M., and 90% at 9:00 A.M It should be noted that at no time during the day were 100% of the staff engaged in work, as might be expected A similar staggered distribution of departure times was used for the 4:00, 5:00, and 6:00 P.M time segments These two work arrival–departure patterns are summarized in Figure 16.10 Note that the staggered hours pattern has more gradual arrival and departure patterns than the relatively rapid transition patterns in the existing work pattern Several of the critical corridors of concern in the peak hours, with significant congestion and delay, are Park Avenue, Atherton Street, College Avenue, and Beaver Avenue (Figure 16.1) To evaluate the effectiveness of the staggered work hours scenario, the volumes on these corridors were compared before and after policy implementation However, because these corridors at times have traffic volumes exceeding capacity in the peak hours, the volume changes due only to the staggered hours policy are difficult to determine User equilibrium traffic assignment was used for the base and TDM scenarios This means an equilibrium point is found such that an alternate route will not decrease the travel time between any origin–destination pair However, the relationship between the base and TDM equilibrium points are not linear As less Penn State traffic use congested routes, delays and travel times may improve to the point where other traffic will shift to these roadways from less attractive and congested routes As a result, the volume reductions in the peak hour not equal the volume increases in the “shoulder” hours (6:00 a.m., 8:00 A.M., 4:00 P.M., and 6:00 p.m.) These nonlinear effects are not obvious in reviewing the proposed policy, but the model can be used to predict how traffic patterns will change as the result of a policy scenario such as this one: The changes for the eastbound (EB) and northbound (NB) directions of the corridors in the morning hours (6, 7, and A.M.) are summarized in Figure 16.11, and the changes for the westbound (WB) and southbound (SB) directions of the same roadways for the same period are shown in Figure 16.12 © 2003 CRC Press LLC 100.00 90.00 Percent of PSU staff at Work 80.00 Existing 70.00 Staggered 60.00 50.00 40.00 30.00 20.00 10.00 0.00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00 1:00 2:00 3:00 AM AM AM AM AM AM AM AM PM PM PM PM PM PM PM PM PM PM PM PM AM AM AM AM Time of Day FIGURE 16.10 Existing and staggered hours of staff work patterns 1000 Existing Staggered Existing Staggered Existing Staggered 900 800 700 Vehicles per Hour 600 500 400 300 200 100 Park @ Atherton Park @ Bigler Park @ Fox Hollow Beaver @ Atherton Beaver @ Shortlidge Atherton @ Park FIGURE 16.11 EB and NB volume changes due to staggered hours policy It should be noted that on each of these links, the volumes under the staggered hours scenario fell within the standard deviation range of the mean volume for that hour The impacts of the staggered work hours scenario were not as great as expected, especially in the Park Avenue corridor (see Figure 16.1), which is of particular interest to Penn State because of its proximity to the campus and the long delays and queues in peak hours The peak hour traffic volume decreases are probably not worth the major effort required to implement the policy Significant resistance to this change by the staff members affected would presumably be encountered, and without significantly better travel conditions, the policy would be viewed as ineffective and unnecessary © 2003 CRC Press LLC 1400 Before Staggered Before Staggered Before Staggered 1200 1000 Vehicles per Hour 800 600 400 200 Park @ Atherton Park @ Bigler Park @ Fox Hollow College @ Atherton College @ Shortlidge Atherton @ Park FIGURE 16.12 WB and SB volume changes due to staggered hours policy 16.5 Summary and Next Steps The Centre SIM model represents the first application in which spatial and temporal information about activity participation is used in a more traditional four-step modeling framework The results of this simulation demonstrate the feasibility of incorporating activity data into the traditional UTPS travel demand model The Centre SIM methodology provides the capabilities of more complex models without the need for significant investments in model development and data collection, a critical feature for small regions with limited resources The effectiveness of this model was demonstrated by validating the output to traffic counts in the study area It was also compared to previous models of Centre County The Centre SIM model has expanded abilities in terms of simulating travel throughout the day and testing network and policy scenarios The model has also been shown to better replicate existing traffic patterns than a previous model (AMIS) Therefore, the Centre SIM model is more accurate, in addition to moving toward a more theoretically sound approach Further, the Centre SIM methodology is a unique application of a geography-based travel demand model using activity data, starting first from the spatial distribution of activities This is in contrast to the traditional approach in activity-based modeling, where an individual is assigned a series of activities and later these activities are assigned to locations In addition, the geographic basis of the model provides for simulation of land use policies and scenarios, although these were not demonstrated here The Centre SIM methodology can also accommodate travel viewed either as a derived demand or as an activity to itself This characteristic is important as travel behavior theory continues to be refined The Centre SIM model will not become outdated in the near future due to advances in travel research, so its capabilities can continue to be enhanced and developed Several issues were identified during development of the Centre SIM model that were not addressed or solved in this application These elements would improve the foundation of the model or its modeling capabilities For the Centre SIM model specifically, a representative activity survey of the Centre County population (not just the Penn State population) would provide a better foundation for the activity patterns Additionally, a new survey would support the definition of population segments and the separation of the population into the segments With a more complete data set, persons between 16 and 18 years of age, discounted in the current model, can be simulated in the Centre SIM model This is an activity currently under way for the entire county that will also provide input to a new long-range plan © 2003 CRC Press LLC Another use of the survey data would be to analyze the trip duration data to develop region-specific friction factors Factors based on local characteristics and travel patterns would be a distinct improvement over the generic factors used in this research The temporal distribution of trip durations can also be used to produce friction factors that vary by time of day, similar to the varied network travel times already used in the model For long-range planning purposes, two components may be added to the Centre SIM model to aid in forecasting First, the addition of a sociodemographic microsimulator would provide the ability to simulate the population through 20 or 30 years (see Chapter 14 of this handbook and the Chung (1997) application) The microsimulator can provide the input needed for Centre SIM to model activity and travel demands for future time points Second, although the Centre SIM model already has capabilities for analyzing land use scenarios, an explicit land use simulation tool can be added as a feature to the model Land use changes and development are critical to the growth of a region over time and are directly tied to travel demand and traffic patterns However, current practice in projecting land use consists of trend analysis and the opinions of local planning experts about where developments of various types may occur These methods are subjective and possibly highly unreliable for long-range planning horizons Examples of new simulation models exist (see Chapter 12 on land use models) and should be used for this application Several smallerscale improvements may also be made to the Centre SIM methodology to improve its current accuracy and simulation capabilities; these are reported elsewhere (Kuhnau, 2001) References Alam, S.B., Dynamic Emergency Evacuation Management System Using GIS and Spatio-Temporal Models of Behavior, M.S thesis, Department of Civil and Environmental Engineering, Pennsylvania State University, University Park, 1998 Chung, J.H., Transportation Impact Simulation of Access Management and Other Land Use Policies Using Geographic Information Systems, doctoral dissertation, Department of Civil and Environmental Engineering, Pennsylvania State University, University Park, 1997 Chung, J.H and Goulias, K.G., Access management using GIS and traffic management tools in Pennsylvania, Transp Res Rec., 1551, 114–122, 1996 Goulias, K.G and Marker, J.T., Jr., Procedure for Using the Access Management Impact Simulation (AMIS) Model within the Context of Act 209 and Act 47, final report submitted to PennDOT, PTI 9819, University Park, PA, 1998 Kuhnau, J.L., Activity-Based Travel Demand Modeling within the Urban Transportation Planning System, M.S thesis, Department of Civil and Environmental Engineering, Pennsylvania State University, University Park, 2001 © 2003 CRC Press LLC ... facility and asset inventories and to study their policies a year ahead (e.g., parking changes and signal coordination), and provide data to study operational improvements in public transportation. .. practice in planning for transportation engineers is still the four-step urban transportation planning system (UTPS) model consisting of trip generation, trip distribution, mode choice, and traffic... considerations, and so forth Because these models have little or no behavioral basis, they cannot be used to evaluate the effects of transportation demand management strategies and other programs and small