Abstract: With consideration of proactive attacks,the problem of Third-Party Logistics (3PL) reliable reverse network design was researched.A multi-level optimization model was established: 3PL reverse network structure was optimizes with network design model to minimize the total costs and make the network satisfy the required service level after attacks;the attack strategy was selected with proactive attacks model to destroy the network to the utmost extent;the network-flows decision model was used to optimize the return quantity of recovery product after attacks.According to the characteristic of optimization model,a two-level compact Genetic Algorithm (cGA) and weighted graph based network-flows algorithm were designed.The top-level cGA is used to solve the problem of 3PL reverse network design.The down-level cGA was used to solve the problem of proactive attack,and the weighted graph based network-flows algorithm was used to solve the problem of network flow decision after attacks.Numerical simulation experiments were presented to verify the reasonability of the model as well as the effectiveness of the proposed algorithm.

Key words: third party logistics, reverse logistics, network design, proactive attacks, reliability, compact genetic algorithm

摘要： 在考虑蓄意攻击的情况下,研究第三方物流可靠性逆向网络设计问题,并建立了多层优化模型,包括网络设计模型、蓄意攻击模型、网络流决策模型。网络设计模型用以优化第三方物流逆向网络结构、最小化总物流成本,同时使网络在遭受攻击后满足一定的服务水平;蓄意攻击模型通过选择攻击策略来最大程度地破坏网络;网络流决策模型用以优化被攻击后回收产品的返回量。根据优化模型的特点,设计了两层紧致遗传算法和基于加权图的网络流算法。上层两层紧致遗传算法求解第三方物流逆向网络设计问题;下层两层紧致遗传算法求解蓄意攻击问题;基于加权图的网络流算法求解攻击后的网络流决策问题。通过仿真实验验证了模型的合理性和算法的有效性。

关键词: 第三方物流, 逆向物流, 网络设计, 蓄意攻击, 可靠性, 紧致遗传算法