Collaborative Optimization For Demand-driven Train Timetable And Train Connection Plan In Urban Rail System

With the acceleration of the urbanization process in China,there is an increasing travel demand for urban transportation,and the contradiction between supply and demand is particularly prominent.Urban rail transits with large capacity,high frequency and reliability become one of the effective way to ease traffic congestion.However,most metro companies still determine the operation plans in accordance to their experiences and professional judgments,causing the waste of energy and absence of accurate computation and optimization.In this thesis,based on careful analysis of the results of existed studies,considering the time-dependent passenger demand and train traffic dynamic,we mainly research the collaborative optimization of train timetable and train connection plan in urban rail system,and the contents are as follows:(1)This thesis formulates a nonlinear optimization model with multi-objectives for the metro train scheduling problem with time-dependent demand,and designs an approximate dynamic programming(ADP)approach.Under the systematic safety constraints,the train traffic model and train energy consumption model are proposed by considering the train headway and train passenger loads as state variables.A nonlinear dynamic programming model for train timetable problem is formulated to generate a near optimal timetable to realize the tradeoff among the utilization of trains,passenger waiting time,service levels,and energy consumption.To overcome the curse of dimensionality in this optimization problem,we construct an approximate dynamic programming framework,where the original problem is first transformed into a Markov dynamic process and then solved by a value function approximation based ADP algorithm.Moreover,this algorithm is able to converge to a good solution with a short time compared to the genetic algorithm and differential evolution algorithm.In addition,the proposed model and algorithm is able to balance those objective functions,and the optimization preference on service quality or operation costs can be adjusted based on practical requirements.(2)Based on the train schedule optimization,this paper presents a collaborative optimization method for metro train scheduling and train connections combined with passenger control strategy on a bi-directional metro line.Specifically,by regarding the train headway,number of stranded passengers at platforms and train passenger loads along the metro line as state variables,and considering the turnaround operations and the entering/exiting depot operations simultaneously,this thesis formulates the collaborative optimization problem as a mixed integer nonlinear programming model to realize the trade-off among the utilization of trains,passenger flow control strategy and the number of awaiting passengers at platforms.Then this model is further reformulated into a mixed integer linear programming(MILP)model by linearizing the nonlinear constraints.To handle the complexity of this MILP model,a Lagrangian relaxation-based approach is designed to decompose the original problem into a train scheduling sub-problem and a train connection sub-problem,which reduces the computational burden of the original problem and can efficiently find a good solution of the train schedule and train connections problem combined with passenger flow control strategy,of which the gap is less than 4%.Under the proposed collaborative optimization approach,the number of train service connections and the crowding inside stations and carriages with the proper passenger flow control strategy can be evidently balanced,and thereby the operation efficiency and safety of the metro lines are effectively improved.