不确定环境下两阶段应急供应链网络建模与优化求解

Modeling and Optimization Solution of Two-stage Emergency Supply Chain Network under Uncertain Environment

为制订有效的应急供应链网络规划方案,提升应急组织救援效率、实现资源物质合理配置,考虑供应和需求的不确定性,采用多运输方式联合的配送模式,以最小化网络响应时间、成本和碳排为优化目标,构建两阶段应急供应链混合整数规划模型,同时,基于鲁棒优化理论构建可调节鲁棒优化模型,增强网络应对不确定因素的能力,通过线性对偶理论对含不确定参数的约束进行转化;为提升模型的求解效果,提出一种优化布谷鸟搜索(optimize cuckoo search,OCS)算法,引入基准实例,以验证OCS算法求解多目标函数的优越性和适用性;最后,利用武汉新冠疫期期间应急物资配送数据,研究带有不确定参数的应急供应链网络决策问题,并通过敏感性分析证明鲁棒控制系数对不确定扰动的有效抑制作用。

In recent decades,all kinds of natural disasters and public health emergencies have occurred frequently.In order to reduce the loss and casualties caused by emergencies as much as possible,emergency supplies must be delivered to all the points of need in the shortest possible time.Therefore,the center location and material distribution in the emergency network become the key problems to be solved after an emergency.To improve the operational efficiency and reliability of the emergency supply chain network,the location and distribution of the emergency supply chain network are studied from the perspective of system integration and optimization.The supply chain network studied includes the supply point of emergency supplies,the transit warehouse of supplies and the demand point.According to the characteristics of emergencies,considering the uncertainty of emergency supplies supply and demand,a two-stage emergency supply chain network planning model is constructed by adopting the multi-transport mode distribution model.Firstly,based on the robust optimization theory,the uncertain demand and supply are represented as interval data,and the linear duality theory is used to transform the uncertain parameter constraints,and a multi-objective robust optimization model is established,with the network response time,cost and carbon emissions as the minimum optimization objec⁃tives.Secondly,a meta-heuristic algorithm is used to solve the model.Considering the shortcomings of the standard cuckoo algorithm,which uses fixed step size control factor and fixed discovery probability to search the optimal solution,an optimal cuckoo search(OCS)algorithm based on dynamic parameter adjustment strategy is proposed in this paper.OCS algorithm is applied to four test problems,namely DZT1,DZT3,DTLZ2 and DTLZ5,and the optimization results are compared with those of CS algorithm and NSGA-II algorithm to verify the effectiveness of the proposed algorithm.The experimental results show that compared with CS algorithm,the solving ability of OCS algorithm has been significantly improved.The introduction of the dynamic adaptive adjustment strategy can effectively improve the convergence and uniformity of the algorithm.In addition,the experimental results show that the OCS algorithm has a strong competitiveness compared with the mainstream NSGA-II algorithm.Finally,the emergency supply chain network decision-making problem with uncertain parameters is studied by using the emergency material allocation data of the disaster areas of Wuhan.The results show the effectiveness of the multi-objective robust optimization model,and the sensitivity analysis results verify the effective inhibition effect of the robust control coefficient on the uncertain disturbance.This study can provide effective guidance for the construction of emergency supply chain network in emergencies.