Transit Loop Line OD Matix Estimation Method Based On Markov Model And Bayesian Analysis

As a green way of travel,public transportation is increasingly.being used by people.As a common and special line form of bus lines,the loop bus line is worthy of study.OD data of bus lines is the basic data of bus network planning,bus schedule and bus dispatching.Therefore,it is helpful to improve the efficiency of public transit network by making the estimation of OD matrix of public transit link efficient and low cost.First of all,the classification of loop bus line is introduced,which is divided into two types:"discontinuous" loop line and "continuous" loop line.The differences between the tworules are distinguished,and the differences of the OD matrices caused by different rules are analyzed.Then the paper introduces the data which is needed to build the model,the data is only the number of passengers on and off the station,there are many advantages such as:single type,simple,high accuracy etc.This paper establishes a first-order Markov chain model,which effectively describes the process of passengers getting on and off.And the unknown parameter in the OD estimation process is reduced to only one variable.This avoids the problem that the OD estimation model is highly underspecified.Using Bayesian inference solves the unknown parameters in Markov chain model.The priori information is used effectively to make the estimation result more accurate.By comparing the first-order Markov chain model with the maximum entropy method without prior information,the first-order Markov chain model is equivalent to the maximum entropy method without prior information.As the first-order Markov chain model assumptions make the model cannot fully describe the reality,as a supplement,the paper proposed a higher order Markov chain model,and the second-order Markov chain model as an example,the model was established and solved,The higher-order Markov chain model can be obtained by analogy with the second-order Markov chain model.Finally,an example is given to demonstrate the model and algorithms proposed in this paper.An example shows that the proposed Markov chain model and Bayesian inference algorithm can avoid the problem of highly underspecified in OD estimation.The closed-form solution is obtained and the priori information is used effectively to improve the accuracy of the OD estimation of the loop transit line.