14465

1360 دكتري

دكتري

مهندسي صنايع

اصفهان : دانشگاه صنعتي اصفهان

1397

دوازده، 116ص.: مصور، جدول، نمودار

نادر شتاب بوشهري

علي زينل همداني

حسين پورزاهدي، مهدي ابطحي، غلامرضا شيران

1398/01/17

كتابنامه

صنايع و سيستم ها

مهندسي صنايع و سيستم ها

1398/01/19

ID1360 دكتري

007 Origin Destination Matrix Estimation and Observation Using Traffic Sensors in Large Scale NetworksAbstractThe travel demand matrix also known as origin destination matrix OD matrix is an effective and efficient instrumentin urban planning as well as managing traffic Given their nature and extent of operation usually direct methods ofestimating the matrix impose very high costs in terms of both time and human resources Thus over the past threedecades numerous attempts have been made to propose alternative methods of estimating and observing the OD matrix One of these methods is the use of information collected by traffic sensors In this thesis it has been attempted toconsider two problems the first one is OD matrix estimation on congested networks in large scale networks and thesecond one is OD matrix observation on uncongested networks Regarding to OD matrix estimation using traffic countson some links the Bayesian inference approach which is the most popular methods for estimating the OD matrix is used By applying the Bayesian inference approach an algorithm is proposed which aims to find the links whose informationshould be optimized In this algorithm unimportant links are removed using a proposed technique Next the OD matrixis updated via Bayesian inference by relying on the information obtained from observed links To this end an efficientcomputational structure is presented which employs the Bayesian approach while significantly reducing time andmemory requirements without loss of accuracy the reduction is more substantial in large transportation networks Toevaluate the proposed algorithm it was tested on the Sioux Falls network The results indicate that the proposedalgorithm offers adequate accuracy in estimating the OD matrix Finally to demonstrate its applicability to large scalenetworks it was used on the transportation network in the City of Isfahan Results proved that it can estimate OD matrixfor large networks without compromising on accuracy within a reasonable time Regarding to OD matrix observation theobjective of this thesis is to observe OD matrix under two scenarios In the first scenario it is assumed that the trafficnetwork is equipped with path ID sensors In this situation the goal is to determine the optimal number and location ofthese sensors in the network where by applying collected information through these sensors the OD matrix is observed Because path ID sensors are not available in many cities in the second scenario the interview alternative is proposed inorder to observe OD matrix This leads to two problems 1 path complexities preclude enquiring about paths and 2 stopping and interviewing all vehicles causes major disturbance in the network A methodology with several models isproposed to overcome these problems The proposed models are applied to the Nguyen Dupuis transportation network which is a small network and the results are analyzed Furthermore by executing the model on the intercitytransportation network in the Province of Isfahan which is a large real network the efficacy of the proposed models isdemonstrated

نادر شتاب بوشهري

علي زينل همداني

حسين پورزاهدي، مهدي ابطحي، غلامرضا شيران