Detector Placement Model Of OD Demand Estimation Considering Observation Error

OD(Origin-Destination)demand estimation is a long-term topic in the interdisciplinary field of operations research and traffic management.This dissertation considers the impact of data collection errors on the accuracy of OD demand estimation and the corresponding optimal locations of detectors so as to investigate the possible errors in OD demand estimation.Detector location models based on static OD demand estimation and stochastic OD demand estimation under multiple types of observed errors are given.The corresponding solution algorithms and numerical examples are presented to study the effect of observed errors of different locations in the traffic network on the accuracy of OD demand estimation.It is useful to avoid placing detectors with large errors on sensitive road sections.The first chapter introduces the research background and current progress of the OD demand estimation problem and the corresponding detector location problem.Then,the error sources for the OD demand estimation and the practical significance of the corresponding detector location problem are discussed.The issue of observed error on OD demand estimation is also addressed.The research progress on the detector location problem for OD demand estimation with consideration of observed error is presented.In the second chapter,we firstly review the classic static OD demand estimation and stochastic OD demand estimation models,the OD demand estimation error and the detector location models.The main sources of the error in the OD demand estimation problem(errors in traffic flow observation and path choice proportion estimation)are reviewed.The classic MPRE(Maximum Possible Relative Error)model of OD demand estimation error is extended,and the MPRE model for examining the accuracy of static OD demand estimation under the condition of multiple types of observed errors is proposed.Then,it is extended to account for the estimation error for stochastic OD demand estimation.The relevant mathematical properties are proved.The third chapter presents a detector location model for static OD demand estimation problem with consideration of different types of observed errors.Numerical examples are carried out to show the influence of different types of observed errors on the OD demand estimation accuracy.Also,sensitivity tests on the locations of observed links and OD pairs are given.In Chapter 4,the proposed detector location model in Chapter 3 is extended to consider the stochastic OD demand estimation problem.The concept of maximum possible relative error is extended into the stochastic OD demand estimation problem with multiple types of observed errors.Numerical examples are carried out to show the influence of different types of observed errors on the OD demand estimation accuracy in a small-scale network example.The efficiency of the proposed algorithm is tested in a medium-size network.The fifth chapter gives the summary,prospects and innovations of this dissertation.This dissertation has 16 pictures,16 tables,and 53 references.