The Research On Route Generation Model Of Real-time Custom Bus Based On Random User's Demand

In the process of the development of public transportation,the traditional bus has been played an important role,vigorously develop public transportation to improve urban traffic congestion,environmental pollution and so on "urban disease" has an obvious effect.But most cities in our country in the process of outward development,without proper planning,led to the outskirts of the city in many large residential community,"separation" post live phenomenon is obvious,but often the peripheral parts of the city bus lines coverage is low,residents travel demand do not match with bus service,need to increase transit coverage in these areas to meet the needs of residents daily travel.In these cities,on the other hand,the peripheral zone due to the low passenger density,fixed-point alignment of conventional public attendance is not high,poor public transport enterprise operating efficiency,severe resource waste.In general,the residents travel demand is all-weather distribution,randomness,and need a few days in advance to make an appointment and only serve the commuter time custom bus often cannot well agree with the characteristics of residents.Real-time custom bus based on demand response is able to solve the low density of the random traffic area passenger travel demand,this article will revolve around its lines generated issues related research.This article first analyzed the key elements of real-time custom bus,including the response patterns,relaxation time window,scope of services,service conditions and scheduling strategy,describes the real-time custom bus service process.Then start time is bounded to the passengers in the station waiting time,passengers in the car waiting time,vehicle running benefit as the optimization goal,the establishment of a two-stage real-time customization model which generates the bus lines,the first stage is the model of vehicle routing problem with time Windows,the second phase multi-goal optimization model.Then according to model characteristics,designed the model algorithm,for the first phase of the OVRPTW model,an improved genetic algorithm to solve,for the second stage of the multi-objective optimization model,the objective weighting method is transformed into single objective optimization problem,using the first stage design improvement algorithm to solve.Final design test cases,randomly generated in an area several dynamic driving request,through the model calculation,respectively to generate the two phases of vehicle real-time optimal path,and analyzes each stage of the passengers in the station waiting time and passengers go time,gives a vehicle arriving time window of each site,through the time window,you can see that the real-time customization model which generates the bus lines at the early stage of the vehicle can well satisfy the demands of passenger travel time,the vehicle running late,due to previous rerouting traffic service,lead to early dynamic ride requests of passengers being late service,vehicle arrives site time outside the passenger waiting time window.