铁路客运票额预售控制决策模型研究  

Control and Decision Model for Railway Passenger Ticket Pre-sale

作　　者：刘帆洨 彭其渊[1,2] 梁宏斌[1] 杨奎[1] 

机构地区：[1]西南交通大学交通运输与物流学院,四川成都610031 [2]西南交通大学综合交通运输智能化国家地方联合工程实验室,四川成都610031

出　　处：《铁道学报》2018年第1期17-23,共7页Journal of the China Railway Society

基　　金：国家自然科学基金(61571375);中国铁路总公司科技研究开发计划(2016X008-J)

摘　　要：客票发售是列车能力利用的具体表现,票额出售即为能力占用,预售期购票需求的变化直接影响着列车能力的占用过程。本文在已知基于列车沿途停靠站的票额分配计划下,以客流需求预测值和最低票额保护值为约束。考虑预售期购票趋势与不同OD旅客的平均购票强度等因素,利用半马尔可夫决策过程来描述票额预售过程,提出单次决策期望收益模型,并以此推广至列车全程及整个预售期,最终构建以最大期望票价收益为目标的单列车票额预售控制决策模型。通过算例对模型进行了验证,结果表明本文提出的方法在考虑旅客需求的同时,可改善列车能力利用,提高列车客座率与收益。Ticket sale reflects the capacity utilization of railway transportation.A sold ticket represents an occupied seat.In the whole pre-sale period,the change of purchase demand directly affects the occupation process of the train capacity.This research is under the given train tickets allocation plan for stations with restrictions of passenger demand forecast and the lowest preserved number of tickets.A Semi-Markov Decision Process was used to describe pre-sale process considering the pre-purchasing trend and the average purchasing intensity for different OD(Origin-Destination)passengers.An expected revenue model was proposed for a single decision and extended to the whole pre-sale process and whole stations along the line for a single train.Then a pre-sale control and decision model was constructed for maximizing the expected fare revenue.The model was verified by a numerical example.The results show that the model can improve the train capacity utilization and increase the passenger load factor and the revenue while it considers the passenger demand.

关 键 词：铁路客运 票额预售 控制决策 半马尔可夫决策过程 最大期望票价收益 

分 类 号：U293.22[交通运输工程—交通运输规划与管理]