Abstract: To solve the capacity design and resource allocation problem at third-party warehousing under uncertain demands, a stochastic mathematical programming model is proposed. It is proved in this paper there exist optimal resource allocation quantities, and the optimal quantity of each class of customer is a decreasing function of unit resource cost, while an increasing function of unit resource revenue and unit lost cost. Because of the complexity for this problem to be solved analytically, a new approach integrating revenue management with discrete-event simulation and response surface methodology is developed. Revenue management is used to implement customer segmentation and construction of resource allocation policies. Simulation is employed to model the stochastic characteristics and evaluate the performance of the system. The optimization of allocation policies and exploration of performance improvement is realized by a statistical experimental technology, response surface methodology. Case study and simulation experiments show that allocating warehousing capacity according to customer segmentation is better than the traditional policy of first come first service, and the combination of revenue management, response surface methodology and simulation can effectively improve system's yield, which validates the approach proposed in this paper.

Key words: third-party logistics, recourse allocation, revenue management, simulation modeling, response surface methodology

摘要： 针对不确定市场需求条件下第三方仓储资源的能力规划与分配问题,构建随机数学规划模型,理论分析证明了最优资源分配量的存在性,并指出最优资源分配量是单位资源成本的递减函数、单位资源收益和单位损失成本的递增函数。鉴于解析求解的复杂性,基于收益管理思想,结合离散事件仿真技术和响应曲面法,提出一种新的分析求解框架:收益管理用于细分顾客、构建资源分配策略,仿真模型刻画系统随机特性并评估系统绩效指标,响应曲面法则优化分配策略并探寻绩效改进方向。案例研究和仿真实验结果显示,根据顾客类别分配仓储能力的策略优于传统的先到先服务策略,收益管理、响应曲面法与仿真的综合集成,能够提高系统收益,从而使本文所提方法体系得到了有效验证。

关键词: 第三方物流, 资源分配, 收益管理, 仿真建模, 响应曲面法