Research On Coupling Optimization Of Arrival And Departure Track Scheduling For Railway Large-scale Passenger Station And Its Receding Horizon Adjustment

The reception and departure tracks distribution plan and its dynamic adjustment are thecore link of large-scale railway passenger station technical operation plan, which needs toconsider coupling relationship among influencing factors that are passenger flow, operationroute at throat, the utilization mode of passenger train sets and so on, and which is a largescale combinatorial optimization problems with many characteristics, such as complexconstraints and multi-objective. To achieve the automatic compilation and dynamicadjustment for the reception and departure tracks’ distribution plan directly affectsorganization efficiency and operation ability of large-scale railway passenger station, and isthe key link and theoretical difficulty of intelligent dispatching system of large-scale railwaypassenger station.Based on references to existing related research achievements, this paper summarizes theresearch situation and development trend of the reception and departure tracks distributionoptimization. With carrying on predecessors achievements, this paper objectively analyzesthe lacks of the existing research and the key problem needed to be solved urgently. Throughthe analysis of the operation organization process, and the compiling basis and content ofdaily (class) plan and stage plan of the large scale railway passenger station, this paperanalyzes the main factors and their intrinsic coupling mechanism affected the reception anddeparture tracks distribution, and the internal mechanism that the passenger train delays ordelay spread affect the reception and departure tracks distribution.Based on this, taking the delivery train sets operating time as originating time and final stoptime of passenger train, respectively. According to this, it can obtain the time window of anypassenger trains occupying the reception and departure tracks, and operation path at thethroat area, and according to the main factors influenced the reception and departure tracksdistribution, and inherent mechanism of pursuing the optimal objective function, as well asconsidering the basic constraint conditions, the coupling constraints between the receptionand departure tracks distribution and route assignment, the coupling constraints between thereception and departure tracks distribution and the passenger flow, and the couplingconstraints between the reception and departure tracks distribution and train sets assignment,and setting the balanced use of all reception and departure tracks, the convenience for thepassenger on or off as the first optimization goal, and setting the minimum cross interferencebetween operation routes when receiving and sending passenger trains as the secondoptimization goal, and finally the multi-objective optimization model (Model Ⅰ) isestablished for the reception and departure tracks distribution plan. Considering the influence that the number of decision variables affects the complexity ofthe optimization model, by introducing "virtual the reception and departure tracks", this paperputs forward to an improved optimization mode（lImproved-ModelⅠ）based on the virtual thereception and departure tracks, which makes the reception and departure tracks distributionand route assignment achieve a further organic coupling, simplifies the complexity of themodel, and makes the optimization model more operational. At the same time, considering therobustness（anti-jamming ability）, which is one of the main problems of the reception anddeparture tracks at the large-scale railway passenger station, analyzing the minimum timeintervals value between two adjacent trains occupied the same reception and departure track,and put forward a method of taking value by the buffer time. The reception and departuretracks distribution optimization models established, not only consider technical operationcoordination of large scale railway passenger station, the interests between passengers and thelarge scale railway passenger station, but also consider coupling relationship between thereception and departure tracks distribution and the utilization mode and train sets.Without considering the random factors disturbance, stage plan of large railway passengerstation arranges the reception and departure tracks according to the established daily (class)plan, this kind of the reception and departure tracks distribution can be regarded as a staticscheduling problem. But because the railway transportation system is a complex, open andlarge system, which is easy to be disturbed by random factors, such as equipment temporaryfault, emergencies, natural disasters and so on, and will inevitably result in the actualexecution plan deviating from the established transportation plan, and directly affect therailway passenger station operation organization order, leading to adjusting the originaloperation plan, and this kind of the reception and departure tracks distribution optimizationcan be regarded as a dynamic scheduling problem. Considering the reception and departuretracks distribution optimization principle, such as short time, less change, and early recovery,setting the minimum degree that all the passenger trains practical reception and departure timedeviates from their scheduling time as the first optimization goal, and setting the minimumnumber of passenger trains that change the occupation between the reception and departuretracks and the reception route as the second optimization objective, a multi-objectiveoptimization model (ModelⅡ) is established to adjust the reception and departure tracksdistribution plan, making the final adjustment keep better performance, besides maintainingthe stability of passenger transportation organization.On this basis, taking into account the problems that the global static optimization is oflarge scale, and slow response speed, according to the daily operation plan following the"planning and plan adjustment" mode, the Rolling Horizon theory and method are introducedinto the reception and departure tracks distribution plan adjustment at the large railway passenger station, and this paper puts forward the real-time dynamic adjustment strategy ofthe reception and departure tracks distribution plan based on the rolling horizon. A longerduration real-time adjustment problem of the reception and departure tracks distribution isdivided into several shorter time periods to deal, thus the solution scale of sub problem isreduced, and a series of dynamic dispatching scheme, which reflect the outside environmentchange, can be obtained. The reception and departure tracks dynamic adjustment strategybased on the Rolling Horizon, can not only overcome the disturbance from the random factorsto the reception and departure tracks distribution plan, but also make the adjustment schemedeviate from the original plan as small as possible. The optimization results are alsoconsistent with the actual demand for the dynamic adjustment of the reception and departuretracks distribution plan, which provide a new way to solve the problem of dynamicadjustment for the reception and departure tracks distribution plan at the large railwaypassenger station.Taking into account the advantages that the simulated annealing algorithm is able to jumpout of local optimum, fast convergence speed, using simulated annealing algorithm and thegeneral theory and methods of multi-objective optimization theory, and using the fast nondominated sorting and density comparison operator and elitist strategy, the simulatedannealing algorithm for solving multi-objective optimization model based on elite strategy isproposed. Finally, setting the LanZhou Railway Station as an example, using the optimizationmodel and algorithm proposed in this paper, calculating by Matlab programming, and throughthe comprehensive analysis of the calculation results, the models are proved to be scientific,reasonable, and feasible. The analysis results show that the reception and departure tracksdistribution plan optimization model, the plan adjustment optimization model and the timeadjustment strategy are scientific and reasonable. Optimization algorithm can obtain multiplePareto optimal solutions a time, which provide decision makers of the large railway passengerstation with multiple choices.