考虑多重不确定性的托管药房物流调度优化

为有效控制药房托管的药品物流运营成本,本文首先结合时空网络架构,将托管药房的药品物流调度过程描绘成一种多层时空网络,继而将研究问题构建为优化领域中的多重货物网络流问题.在此基础上,综合考虑药品物流调度过程中的需求不确定性、配送服务时间不确定性以及药品供应成本的可变性,建立基于药房托管模式的药品物流调度随机优化模型并设计具有挑选规则的混合遗传算法进行求解.参数测试表明,当挑选判断系数设为0.6时,求解算法可以收敛到比较理想的解;算例分析表明,本文所设计的随机优化模型在各决策回合的解与完全信息模型解之间的逼近程度不一,但总误差仅为0.56%;敏感性分析则发现药房最低安全库设置存在一个最优均衡点.

Optimization of logistics scheduling for hospital pharmacy trusteeship under hybrid uncertainty

This paper proposes an effective method to control the operation cost of medicine logistics activities and to improve the robustness of medicine logistics scheduling for hospital pharmacy trusteeship. Initially, medicine logistics activities in the hospital's pharmacy is constructed as a multi-layer time-space network and then the problem is formulated as a multi-commodities flow problem. Considering uncertain demand in each node, uncertain service time between each two nodes and the variable service cost, a stochastic programming model is presented for the medicine logistics planning with the background of hospital pharmacy trusteeship. A hybrid genetic algorithm which compared with a special selection rule is designed for solving the optimization model. Also, an evaluation method is presented to evaluate the performance of the proposed model. The test results show that the optimized result will be obtained when the selection coefficient is set to be 0.6. Although the gaps between the two results in each planning cycle are different, the gap for the whole planning time is 0.56% only. Meanwhile, sensitivity analysis shows that there will be an optimal break-even point for the safety stock setting.