考虑通售席位的旅客列车票额优化方法

摘要

This paper studied the seat allocation method,under the assumption of stochastic demand,for pas-senger trains that considered First-Come-First-Serve (FCFS)seats.During the organization of railway seat al-location,the fixed number of seat inventory could not only be allocated to meet the demand from different OD (Origin-Destination)stations,but also be set as FCFS seats that can accommodate redundant demand when OD tickets are sold out.Due to stochasticity in ticketing process and the nature of FCFS,the solution to the seat allocation model can not be obtained through analytic method.Therefore,the problem was first simplified to a deterministic linear programming (DLP)model to generate initial feasible solution.Then,through random sampling to simulate the ticketing process,with an estimation of sample yield gradient,the solution was re-peatedly updated based on gradient descend method in iterations.Through experimental tests,the revenue by the proposed model was increased compared to DLP solution.%研究随机需求下,考虑通售席位的铁路票额分配方法.售票组织中,有限的席位能力不仅可以分配至不同的OD对(上车站、下车站),还可设置为通售席位,以供其他OD当已分配的票额不足时使用.由于购票过程具有随机性和通售席位"先到先得"的特征,使考虑通售席位的票额分配模型难以用解析方法直接求解.因此,首先将问题简化为确定性线性规划模型求得初始可行解,然后对随机样本采样,模拟售票过程,近似计算样本收益梯度,从而迭代优化决策变量.实例表明,对比确定性线性规划模型的分配结果,本文的票额优化方法使收益有所提高.