Timetabling Schemes For Demand Responsive Public Transportation System

With the rapid development of real estate industry in China in recent years,a large number of residential areas have appeared around large cities.But there are fewer commercial circles and service facilities near these residential areas,so residents still need to work,go to school and shop in the central city.These rigid demands have brought travel demand to and from the central city and surrounding suburbs.However,due to the excessive dispersion of residential areas,the city’s bus system model does not serve these areas well.Therefore,the public transportation system of low-demand corridors has a mismatch between supply and demand.In the long run,after the economic level of these surrounding suburbs is raised,there will inevitably emerge the same high travel demand as the central city,and it is also important to match the supply and demand of highdemand corridors.Based on the above problems,this paper hopes to balance the supply and demand relationship of public transportation system through detailed analysis of the characteristics of high and low demand corridors.The research results and main contributions of this paper are as follows:1.Designing bus timetabling scheme by analyzing the characteristics of low demand corridor.Any station in timetabling scheme can be used as the originating station and terminal station of the bus,so it includes the existing three bus system operation strategies(short-turning,deadheading and limited-stop).This timetabling scheme is demand responsive.2.Based on passengers’ choice in low demand corridor,the passenger selection behavior and bus operation process are transformed into a time-space network with cost function and the time-space network model is used to assign passengers.The time-space network consists of nodes and arcs,each of which has its own arc cost function.The distribution result of time-space network model ensures that each passenger selects the bus with the lowest travel cost.3.Design a bi-level model for the low demand bus system to solve the timetabling scheme.The upper layer of the model uses the LNS-based algorithm RRLNS to iteratively update the timetabling scheme,and takes the upper-layer optimized timetabling scheme as input to the lower layer.The lower layer of the model uses the time-space network model to assign passengers to calculate the total cost of the system,and adjusts the optimization direction of the upper layer according to the total cost.4.Considering that the number of decision variables of the subway system is too large if the refined departure intervals are used,this paper designs the analytical algorithm ADAS to solve the subway timetable.The algorithm utilizes the relationship between the service capability of the subway system and the passenger demand,and solves the timetable by analytical solution,avoiding solves large-scale integer programming problems directly.5.The feasibility and robustness of RRLNS are illustrated by the examples,and the optimization direction of the algorithm is verified to be the same as expected.The subway simulation system is built,which is based on the transfer number of passengers between stations.Five constraints are set up to simulate the actual subway system operation process.The feasibility and robustness of ADAS are illustrated by the examples.The ADAS timetable of Beijing Metro Line 15 is obtained,and the application value of ADAS for the subway operator is analyzed.