基于改进混合蛙跳算法的多约束车辆路径优化

针对多中心分布式企业存在的产品成本差异化问题，建立包括产品成本、多车场、多车型在内的多约束车辆路径模型，并设计求解该模型的改进混合蛙跳算法. 根据问题特性，改进聚类算法并结合邻近矩阵构造初始青蛙种群；提出子群概念，设计自内而外的交流演化模式；定义远离矩阵，对青蛙进行引导性邻域搜索. 将所设计的算法进行多组不同的对比实验，结果表明，所设计的算法通用性强，实用性高，与遗传算法、蚁群算法这类传统经典算法相比，具有更好的收敛速度与求解精度，可以有效解决此类问题；考虑产品成本的调度方案总成本平均减少6%，占产品总成本的13%，可以为企业提供更合理的车辆配送方案.

Multi-constrained vehicle routing optimization based on improved hybrid shuffled frog leaping algorithm

Aiming at the product cost differentiation problem of multi-center distributed enterprises, a multi-constrained vehicle routing model including product cost, multi-depot and heterogeneous vehicle was established, and an improved hybrid shuffled frog leaping algorithm was designed to solve the problem. The initial frog population was constructed by improved clustering algorithm and adjacency matrix according to the characteristics of the problem. The concept of subgroup was proposed to design the evolution model of communication from inside to outside. The guided local search was performed in the subgroups according to the distance matrix. The designed algorithm was subjected to many different sets of comparative experiments. Results show that the designed algorithm is highly versatile and practical. Compared with traditional classical algorithms such as genetic algorithm and ant colony algorithm, it has better convergence speed and accuracy, which can effectively solve the problems. The total cost of the solution considering the product cost was reduced by an average of 6%, accounting for 13% of the total product cost, which can provide more reasonable vehicle distribution plan for enterprises.