混合蝙蝠算法在包装废弃物逆向物流中的应用  

Application of Hybrid Bat Algorithm in Packaging Waste Reverse Logistics

作　　者：彭维[1] 朱云波[1] PENG Wei;ZHU Yun-bo(Chongqing City Management College,Chongqing 401331,China)

机构地区：[1]重庆城市管理职业学院,重庆401331

出　　处：《包装工程》2019年第1期253-258,共6页Packaging Engineering

基　　金：重庆市教育委员会2017年度科学技术研究项目(1609155488)

摘　　要：目的为了提高蝙蝠算法(BA)求解包装废弃物逆向物流问题的性能。方法在标准BA算法的基础上提出混合蝙蝠算法(HBA)。首先,构建新型蝙蝠表达式,使BA算法适用于包装废弃物逆向物流问题的求解。其次,引入自适应惯性权重,改造蝙蝠速度更新公式;然后,引入粒子群算法(PSO),对每次迭代中任一随机蝙蝠进行粒子群操作;最后,利用HBA算法对企业实例和标准算例进行仿真测试。结果企业最优回收距离为776.63 km。与遗传算法(GA)、蚁群算法(ACO)和禁忌搜索算法(TS)相比,HBA算法能够求得已知最优解的标准算例个数最多为6个,求得最好解与已知最优解的平均误差最小为8.58%,平均运行时间最短为4.39s。结论 HBA算法的全局寻优能力、稳定性和运行速度均优于GA算法、ACO算法和TS算法。The work aims to improve the performance of the bat algorithm(BA) used in solving the problem of packaging waste reverse logistics. Based on standard BA, a hybrid bat algorithm(HBA) was proposed. Firstly, a new bat expression was constructed to make BA suitable for the solution to the problem of packaging waste reverse logistics. Secondly, the speed update formula of bat was reconstructed by introducing the adaptive inertia weight. Then, particle swarm optimization algorithm(PSO) was introduced, which carried out particle swarm operation on any random bat in each iteration. Finally, simulation tests were carried out on enterprise examples and standard examples by HBA. The optimal recovery distance of enterprise was 776.63 km. HBA could get better results compared with genetic algorithm(GA), ant colony optimization algorithm(ACO) and tabu search algorithm(TS); such as, the number of standard examples of the known optimal solution was 6, the average error between the best solution and known optimal solution was 8.58%, and the minimum average running time was 4.39 s. HBA algorithm is superior to GA, ACO and TS in global optimization ability, stability and running speed.

关 键 词：包装废弃物 逆向物流 蝙蝠算法 粒子群算法 

分 类 号：X79[环境科学与工程—环境工程]