Forecasting And Application Of Passenger Number At Bus Stops Based On Markov Chain

With the deepening of the process of public transport network,the number of bus stops and modes of public transport that passengers can choose gradually increases,and the satisfaction of different bus stops to passenger demand is also different.This paper includes four models:The first is to establish a discrete-time Markov chain prediction model with departure time as time parameter;The second is to divide the early rush hours and establish a continuous-time mean Markov chain model;The third is to establish a continuous-time prediction model based on clustering Markov chain.By forecasting the passengers at the stop,the passengers at the stop can be obtained.Fourthly,a multi-objective optimization model of vehicle parking time interval is established based on the passenger prediction of the stopping station to optimize and analyze the stopping time interval in the transit line.The main contents include:1.By analyzing the characteristics of passenger demand at the stop,setting discrete and continuous time conditions,three different passenger number prediction models based on Markov chain are constructed,which are composed of state transition probability matrix,and an evaluation index composed of error analysis is set up.2.Analyse the adaptability of different forecasting models in the forecasting process,collect the basic passenger data of Changsha Zhongyi road stopping station according to different time conditions,solve the forecast value of passenger number through the established model,and analyze the error according to the evaluation index,and get the adaptive results of the three models.3.On the basis of passenger number prediction,the optimization of bus operation cost and passenger waiting cost is transformed into single-objective optimization of total social cost,and the optimization model of bus parking time interval based on passenger number is constructed.Particle swarm optimization algorithm is used to solve the optimization model,and the adaptability of the optimization model is analyzed.