Cognitive Aspects of Timetable Visualization Support Decision Making Procedia Computer Science 103 ( 2017 ) 94 – 99 1877 0509 Published by Elsevier B V This is an open access article under the CC BY N[.]
Available online at www.sciencedirect.com ScienceDirect Procedia Computer Science 103 (2017) 94 – 99 XIIth International Symposium «Intelligent Systems», INTELS’16, 5-7 October 2016, Moscow, Russia Cognitive aspects of timetable visualization: support decision making N.N Klevanskiy*, M.A Antipov, A.A Krasnikov Saratov State Agrarian University named after N.I Vavilov, 1, Teatralnaya square, Saratov 410012, Russia Abstract This paper demonstrates timetabling visualizations – timetables, transport timetables, multi-project schedules These visualizations present an allocation of system resources The visualization of high-school timetable demonstrates a timetable of student groups The visualization of transport timetable shows arrivals/departs across stations and lines The visualization of multi-project schedule includes Gantt chart and an allocation of system resources These visualizations are provoked a new approach to timetable optimization The timetabling problems can be solved efficiently by two-stage algorithm developed in database system The first, a set of demands must be developed as initial timetable A set of local and global resources are available for carrying out the activities of the systems The solutions obtained by the first stage algorithm with the best resource allocation rule are used as a baseline to compare those obtained by the latter The second, the initial timetable must be optimized The basic criterion for optimization operations is demanded as criterion of resource equability The latter is equal a root-meansquare deviation from a middle value © 2017 The Authors.B.V Published byopen Elsevier B.V Published by Elsevier This is an access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the XIIth International Symposium «Intelligent Systems» Peer-review under responsibility of the scientific committee of the XIIth International Symposium “Intelligent Systems” Keywords: cognitive visualization, timetable, transport timetable, multi-project scheduling, root-mean-square deviation, multi-vectorial ranking Introduction Cognitive graphics images allow representing the contents of studied object or process on a computer screen A cognitive graphics image visually and clearly reflects the essence of a complex object, and is also capable of providing a fundamentally new decision Thus, an important feature of cognitive graphics image is targeted influence on the intuitive thinking mechanisms Creation of information technologies, based on the use of cognitive graphics images, implemented in intelligent systems for decision-making of different problems in various concrete * Corresponding author E-mail address: nklevansky@yandex.ru 1877-0509 Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Peer-review under responsibility of the scientific committee of the XIIth International Symposium “Intelligent Systems” doi:10.1016/j.procs.2017.01.020 N.N Klevanskiy et al / Procedia Computer Science 103 (2017) 94 – 99 and interdisciplinary problem areas Cognitive graphics images are used in a variety of in intelligent systems for optimization of the learning process, for visualization and prediction of education process results, etc.nalysis of information structures of knowledge and data, identification of different kinds of regularities in data and knowledge, and decision-making and decision justification, in intelligent training and testing systems for Scheduling is an arrangement of entities into a pattern in space-time in such a way that constraints are satisfied and certain goals are achieved Constructing a schedule is the problem in which time, space and other resources have to be considered in the arrangement The class of scheduling problems includes a wide variety of problems such as multi project scheduling, transport scheduling, educational timetabling and many others Many real world scheduling problems are multiobjective by nature, i.e several objectives should be achieved simultaneously A number of multiobjective metaheuristics have been proposed to obtain solutions for multiobjective optimization problems2 Metaheuristics include tabu search, genetic algorithms, neural networks, ant colony optimization and many others The purpose of the paper is demonstration different timetable visualizations provoking new metaheuristic approach to timetable optimization Timetable visualization techniques Schedule task software solution uses a two-step approach3 - forming primary schedule and its subsequent optimization The initial timetable is any timetable that satisfies the hard constraints Both phases required one visualization image just schedules for qualitative analysis Image schedule represents the allocation of system resources in time within the schedule interval Visualization analysis of the initial schedule allows define approaches to optimization and visualization of the resulting schedule provides an opportunity to evaluate hypotheses put forward A high school timetable distributes three kinds of resources - student groups, teachers and audience Figure presents the results of formative software 927 schedule sessions for 50 student groups4 Figure represents the distribution of a system resource - groups of students within the two-week schedule interval Classes in each group for a "couple" of the two weeks of the schedule are one under the other The color represented various occupations: red - lecture, blue - practical, green - laboratory Compact presentation allows to reach a glance results and generate some conclusions The main consumers of class schedules are students whose initial timetable (Fig 1) is non comfortable Firstly, the number of group lessons on different days of the week varies from one to four Secondly, the start time of the first classes on different days is different Thirdly, the number of lessons the same day in different schedule weeks, too, is different These considerations have made it possible to formulate equability grades of group lessons These grades form the equability criteria classes schedule and strategy optimization of initial schedules4 Figure shows the results of the optimization of the initial schedule Visual comparison shows the removal of most of the listed flaws Mutual permutation classes optimized schedules (Fig 2) the possible solution to secondary tasks satisfying the wishes of teachers and classroom efficiency fund Resource heterogeneity - specificity groups, teachers and audiences prevents the formation of integral assessment of resource allocation in the high school timetable For transport systems as opposed to educational institutions cannot be generating views schedules throughout the system The traffic gives rise to several kinds of schedules for different types of system resources Schedule of the particular vehicle can be represented as a vector of times of arrival/departure stops at various points Transport schedules through a particular stopping point is usually the details view arrival/departure times It is necessary to schedule the movement of vehicles between the separate paragraphs For transport systems formed work schedulesairplane crew, locomotive crews, etc For schedules of stations and hauls between them stopovers used spiral presentation (Fig 3), which is a temporary spiral axis, and its length is equal to the interval schedule Spiral turn - the smallest period schedule Showing spiral possibly representation of arrival/departure of the vehicle In Figure presented the initial schedule of passenger transport via the most loaded station railway network test5 The timetable is shown weekly, and each spiral turn consistently represents the per diem schedules The first internal turn is schedule for Monday Blue presents the arrival/departure to the station of trains, yellow-two trains, while staying at the station In the upper-right corner of the specified value of the root-mean-square (RMS) deviation from the mean interval arrivals/departures at the station 95 96 N.N Klevanskiy et al / Procedia Computer Science 103 (2017) 94 – 99 Fig Initial high-school timetable Analysis of Figure leads to the conclusion that it was necessary to ensure the equitable use of resources station Achieving equitable utilization of resources station may equal intervals the arrival of trains The introduction of the RMS deviation from the mean interval as estimates of uniformity and uniformity criteria station trains to all stations of the route it is possible to choose the most uneven trains5 For this train within days is determined by the time of departure from the initial station using multiple criteria ranking uniformity across all stations and hauls route In Figure submitted programmatically obtained optimized schedule with RMS twice reduced For presentation schedule in the project scheduling and multi- project scheduling in most cases used Gantt charts by which are temporal sequence of executing projects and works of individual projects Depending on the subject area have different schedules for the allocation of resources Fig Optimized high-school timetable For numerical experiments multi-project scheduling used randomly selected from a library of test tasks PSPLib6 N.N Klevanskiy et al / Procedia Computer Science 103 (2017) 94 – 99 15 projects Projects include 30 works and they need resource type Figure shows the initial multi-project schedule with the adopted schedule interval and taken project aggregations7 At the top of the picture is the Gantt chart for the 15 selected projects when adopted schedule interval to 100 cycles of planning At the bottom of the Figure chart showing consumption (blue) and distribution (sky blue) for each of the four resources at the each planning cycle Digits in charts resources represented the maximum values of the clock of consumption Red lines demonstrate average consumption of a resource in an interval schedule The second digit indicates the root-mean-square deviation in % from the mean value, shown in red Fig Initial timetable of the most loaded station Fig Optimized timetable of the most loaded station For making decisions about the optimization strategies of primary schedule (Fig 5) used the fact that significant 97 98 N.N Klevanskiy et al / Procedia Computer Science 103 (2017) 94 – 99 resource consumption inequality within the schedule interval Economic efficiency requires the need to achieve a uniform resource consumption in interval schedules This leads to uniformity of schedule for each resource-mean square deviation from the average consumption of a resource in an interval schedule Vector of four evaluations of resources forms schedule irregularity criterion Identifying the most uneven of the project and find the best time for him to start within the schedule interval by ranking the uniformity criterion is the strategy of each cycle of optimization8 Figure presents the optimized schedule Fig Initial multi-project schedule Fig Optimized multi-project schedule N.N Klevanskiy et al / Procedia Computer Science 103 (2017) 94 – 99 Conclusions The authors believe that the new provisions are the following: x developed various kinds of visualization schedule forms submitted system resources consumed; x formulated common approaches to optimize initial schedules based on minimization of RMS deviations from average levels of system resources consumed in schedule interval; x confirmed the validity of the concept of equability of system resource consumption; x presented the results of the optimization initial schedules of various types References Zenkin AA Cognitive computer graphics (in Russian) Moscow: Nauka; 1991 Lazarev АА, Gafarov ER Scheduling theory: problems and algorithms (in Russian) Moscow: Physics Department MSU; 2011 Klevanskiy NN Basic concepts of scheduling problems (in Russian) Educational Resources and Technologies 2014;2(5):9-21 Klevanskiy NN High school and university timetabling (in Russian) Educational Resources and Technologies 2015;1:34-44 Klevanskiy NN, Kravtsov EF Mathematical modelling of initial multi-period timetables (in Russian) Bulletin of Saratov State Technical University 2009;3:100-6 Kolish R, Sprecher A PSPLIB - A project scheduling library Eur J Oper Res 1996;96:205-16 Klevanskiy NN, Krasnikov AA Multi-project scheduling problems methods (in Russian) Educational Resources and Technologies 2015;3:3960 Klevanskiy NN Ranking methods for timetabling problems (in Russian) Proc of XII All-Russian Conference on Control Problems 2014; p 8040-50 99 ... process results, etc.nalysis of information structures of knowledge and data, identification of different kinds of regularities in data and knowledge, and decision- making and decision justification,... shown in red Fig Initial timetable of the most loaded station Fig Optimized timetable of the most loaded station For making decisions about the optimization strategies of primary schedule (Fig... confirmed the validity of the concept of equability of system resource consumption; x presented the results of the optimization initial schedules of various types References Zenkin AA Cognitive computer