研究了逆向需求动态出现的混合集散货物路线问题。将问题的求解过程分为静态和动态两个阶段,每个阶段分别用记录更新算法求解。记录更新法中采用随机大邻域搜索法及重起和扰动策略来提高解的质量,并通过约束邻域搜索范围和设计常量可行性检查两种策略来减少计算时间。通过算例表明了该算法的有效性,同时也表明将动态出现的逆向需求实时整合到已有路线中,比为动态需求单独安排配送更为经济。
Mixed vehicle routing problem with dynamic reverse demands and time windows was studied, solving process was divided into static and dynamic phases which were respectively solved by record-to-record travel algorithm. Stochastic large neighborhood search, restart and disturbance were adopted to improve solution; restricting neighborhood search and designing constant feasibility check were employed to reduce computation time. The computational results demonstrated the effectiveness of this method and integrating dynamic reverse customers timely in the current solution was more economical than sole arrangement.