求解卸装一体化的车辆路径问题的混合粒子群算法

提出了结合粒子群算法（PSO）和变邻域下降搜索（VND）的混合粒子群算法（PSO-VND），用以解决卸装一体化车辆路径问题（VRPSDP）。在此混合算法的前半部分，运用粒子群算法对解空间进行搜索，找到相对较优的一组解。在PSO过程中对于可行化和优化后的粒子添加速度分量，并依据相似度进行变异。在此算法的后半部分，运用变邻域下降搜索对得到的较优解再进行深度搜索，以得到理想的解。在变邻域下降搜索（VND）过程中使用3种不同的邻域结构：插入、交换和交叉，依次对解进行迭代优化。最终采用标准算例进行了仿真试验，验证了混合算法的可行性和有效性。

A Hybrid PSO Algorithm for Vehicle Routing Problem with Simultaneous Delivery and Pickup

It proposes a hybrid algorithm PSO- VND, which combines Particle Swarm Optimization and Vari- able Neighborhood Descent, and applies this algorithm to solve the vehicle routing problem with simultaneous delivery and pickup. During the PSO procedure, it adds a velocity component to the movement of any particle which has been optimized or made feasible, and mute is adopt to particle accord its level of similarity. During the VND procedure, three different neighborhood structures, insertion, swap and cross are successively used. Computational results about the Min and CMT benchmark problems show that this hybrid heuristic algorithm is effective.