Research On Seat Inventory Control Optimization For High-Speed Railway Based On Multi-Level Fare System

With the construction and development of the high-speed railway in China,High-speed railway has become the main mode of passenger transportation in China.However,the high-speed railway construction investment is large,the operation cost is high,many high-speed railway lines are in a loss state.The optimization of revenue is an urgent problem to be solved.High-speed rail tickets are no longer sold with the fixed fare rate and the single price.Therefore,the multi-level dynamic fare system will be implemented step by step.According to the perspective of optimizing the benefits of high-speed railway,this paper study on the issue of seat inventory control systematically based on the multi-level fare system.Firstly,the paper analyzes the problem of seat inventory control based on the multi-level fare.Then,the approximate solution of the basic model is studied in two aspects.Based on the decomposition of dynamic programming and simulation method,the optimization of seat stock dynamic control is studied.The stochastic programming model is used to approximate the basic model,and the optimization of seat stock control based on reservation limit control is studied.Finally,considering the application of the variable train composition of EMU,the optimization problem of seat allocation based on a certain flexible capacity is studied.The main research contents and conclusions are as follows:(1)The seat inventory control problem under the multi-level fare system is analyzed theoretically and the basic model is constructed.Firstly,the characteristics of high-speed railway operations in China are analyzed,and the feasibility of implementing revenue management is analyzed from the perspectives of the policy environment,technical and economic characteristics,and information technology.Then,based on the balance theory between supply and demand,the problem of seat inventory control based on a multi-level fare system is analyzed.Finally,considering the passenger choice behavior,a dynamic programming model is established for the seat inventory control problem based on the multi-level fare system.Based on the characteristics of this model,ideas of approximately solving the model and generating control strategy are pointed out,which propose a foundation for subsequent research.(2)Dynamic optimization of seat control strategy under a multi-level fare system based on dynamic programming decomposition is studied.In this paper,the dynamic programming decomposition and simulation method are used to approximately solve the seat inventory control model.This paper combines the strategy iteration algorithm based on the least-square method and the category optimization algorithm based on the Markov chain choice model to propose an approximate solution algorithm for seat dynamic control model and algorithm for generating real-time control strategy.The feasibility of the dynamic control mechanism can be verified by the simulation experiment of high-speed railway.(3)The seat allocation model under the multi-level fare system of high-speed railway is proposed based on stochastic programming and control strategies are simulated.In this paper,the stochastic programming is used to study the seat inventory control model.The random demand is estimated considering the passenger choice behavior.The seat inventory allocation model of the high-speed railway with multiple trains and stops is built based on the multi-level fare system.Then the model is transformed into the equivalent linear programming model and solved quickly by ILGO CPLEX.Based on the optimal solution,the single-stage static control strategy of booking limits and the multi-stage dynamic control strategy of booking limits are generated.The simulation algorithms are designed to compare the performance between two control strategies.Finally,the validity of the model is verified by numerical experiments.(4)Seat inventory control optimization based on variable train composition is studied for high-speed railway.The variable train composition technology enables high-speed trains to obtain certain flexibility on seat capacity by adjusting the train composition plan.The thesis focuses on the situation of a single train and multiple trains and considers continuous random demand and discrete random demand,respectively,establishes joint optimization models of variable train composition and seat inventory allocation,and gives the corresponding solution methods.Numerical experiments show that the comprehensive optimization of revenue management and flexible train composition has a good effect.When the demand is small,the flexible train composition helps match supply and demand better,and a higher profit is achieved by reducing operating costs.When demand increases to a certain extent,the profit can be increased by balancing the costs and revenue.The main innovations of the paper are as follows:(1)An optimization method for seat dynamic control of high-speed railway is proposed,comprehensively considering random demand,passenger choice behavior,multi-level fare system,and multiple trains and stops.This thesis designs a two-stage control mechanism for the high-speed rail seat inventory control model based on the multi-level fare system,and proposes an approximate solution algorithm that combines simulation-based approximate dynamic programming algorithm with Markov chain category optimization algorithm and algorithm generating online real-time control strategy,and provides a new solution for customer choice-based network revenue management problem.Compared with the existing research results,the algorithm in this paper can obtain the real-time dynamic optimization strategy of high-speed railway seat control.This method can be applied to a wide range of choice behavior models and has good scalability.When sufficient historical sample data is accumulated,the techniques of big data analysis and machine learning can be applied to periodically update the choice behavior parameters and bid price to achieve continuous optimization of the control strategy.(2)A static seat inventory allocation model of high-speed railway is proposed,which considers stochastic demand,passenger choice behavior,multi-level fare system,multiple trains and stops,and the single-stage and multi-stage seat stock control strategies are generated.In the existing researches,the distribution of high-speed railway seats is mainly concentrated in the seat allocation of a single train or multiple trains between different ODs under a single fare system.A small number of seat allocation studies under multi-level fare are based on the assumption of independent demand.To optimize the high-speed rail revenue,this paper uses the stochastic programming method to approximate the high-speed rail seat inventory control model based on the multi-level fare system and constructs the high-speed rail seat inventory allocation model based on the passenger choice behavior.A method of collaborative seat inventory allocation of multiple high-speed trains under a multi-level fare system is proposed.Based on the optimal allocation results,a single-stage static seat stock control strategy and a multi-stage dynamic seat stock control strategy are generated.(3)A seat allocation optimization model of high-speed railway based on variable train composition is proposed.The seat inventory control problem under the condition of a variable train composition plan has not been studied in the existing literature.The existing theoretical research on revenue management mainly focuses on fixed capacity,and a few theoretical research findings on flexible capacity are in the context of aviation aircraft swapping.In this paper,based on the adjustment of seat capacity by variable train composition strategy,the comprehensive optimization model of variable train composition plan and seat inventory allocation is proposed.Some basic rules of seat inventory allocation under variable train composition are obtained through experimental analysis.