论文部分内容阅读
在离散随机需求情景及概率不确定条件下,针对风险厌恶的库存管理者,建立了基于条件风险值(CVaR)的单周期库存鲁棒优化模型.在仅知离散需求情景条件下,结合统计学理论,采用Ф-散度构建了一定置信水平下的不确定需求概率的置信域;运用拉格朗日对偶理论,将单周期库存鲁棒优化模型转化为易于求解的数学规划问题.特别地,给出了仅知需求情景数据下,基于数据驱动的单周期库存策略.最后,进行了数值计算,分析了不同风险厌恶程度、Ф-函数形式和抽样规模对库存策略和库存管理者绩效的影响.结果表明,基于Ф-散度的鲁棒库存策略具有良好的鲁棒性,能够有效抑制需求概率不确定性对库存绩效的影响.进一步,与数据驱动结果对比,发现基于Ф-散度的鲁棒库存策略能够保证库存管理者获得更为理想的绩效,表明对需求数据所蕴含的统计信息的挖掘能够有效改进库存管理者的运作绩效.
Under the conditions of discrete stochastic demand scenario and probability uncertainty, a inventory optimization model based on conditional VaR (CVaR) for single-period inventory is established for risk averse inventory managers.Under the condition of only discrete demand scenario, Theory, the confidence domain of uncertain demand probability under a certain confidence level is constructed by using Ф-divergence, and the robust optimization model of one-cycle inventory is transformed into a mathematical programming problem that is easy to solve by using Lagrange duality theory.In particular, The single-cycle inventory strategy based on data-driven only under the condition of demand data is given.Finally, numerical calculation is carried out to analyze the effects of different levels of risk aversion, Ф-function and sample size on inventory strategy and inventory manager performance The results show that robust stock strategy based on Ф-divergence has good robustness and can effectively restrain the impact of demand uncertainty on inventory performance.Furthermore, compared with data-driven results, it is found that based on Ф-divergence Robust inventory strategy to ensure that inventory managers achieve better performance, indicating that the demand data contained in the statistical information mining can effectively improve Kept operational performance managers.