交通运输工程与信息学报

随着我国电动汽车市场的不断发展，动力电池的退役处置问题变得刻不容缓。实践表明动力电池的有效回收不仅能够避免退役电池对环境造成不必要的二次污染，也能推进动力电池的循环利用，提高经济效益。为有效降低制造商责任制下的退役动力电池回收物流成本，促进动力电池回收程序规范化、可持续性发展，本文以车辆启用成本、行驶成本及能耗成本为总目标，研究退役动力电池回收的多车场车辆路径问题。首先，构建考虑电动物流车运载、载重影响能耗、多车场协同回收的数学规划模型；然后，设计基于遗传算法和局部搜索相结合的文化基因算法，采用无分隔符的编码方式和splitting解码方式、多种结构各异的交叉算子和局部搜索算子、种群管理等策略，提高算法的求解质量和效率；最后，进行仿真实验评估所设计算法的性能。实验结果表明：所设计算法的求解结果与LINGO软件所求最优解的偏差在1.23%以内，并能够在较短时间内求得问题的高质量解；与传统的遗传算法相比，算法在求解质量和效率方面均更优，实验结果验证了算法的有效性。

With the continuous development of the electric-vehicle market in China, issues related to recycling of power batteries have progressively worsened. In practice, effective recycling of power batteries cannot only avoid unnecessary secondary pollution to the environment, but also promote general recycling of power batteries, thereby leading to economic benefits. To effectively reduce the logistics cost of power battery recycling under the manufacturer's responsibility system, and to promote the standardization and sustainable development of procedures for power battery recycling, in this study we addressed the vehicle routing problem for power battery recycling with the objective of minimizing the total cost,including vehicle fixed costs, driving costs, and energy consumption costs.In this paper, a mathematical programming model that considers electric vehicles, load-impact energy consumption, and multi-depot is first formulated;then, a memetic algorithm(MA) based on a genetic algorithm and local search is proposed, and various strategies such as non-separator encoding and splitting decoding methods, multiple crossover and local search operators, and population management strategies are exploredto improve the effectiveness and efficiency of the algorithm; finally,simulation experiments conducted to evaluate the performance of the designed algorithm are presented.The computational results show that the gap of the best objective values between MA and the software LINGO is less than 1.23%, and the best solution by MA can be obtained within a very short time; the proposed algorithm also outperforms the traditional genetic algorithm in terms of computational time and objective values. The numerical experiments confirm that the proposed algorithm can effectively solve the considered vehicle routing problem for power battery recycling.