Abstract: To solve the problem of corridor layout in manufacturing and service system,which assumed that the material loading and unloading points were overlapped and the material flow between facilities was symmetrical,the corridor allocation problem was extended in combination with the actual production layout.To minimize the logistics cost,the corridor allocation problem considering the material loading and unloading points and asymmetric flow was proposed,and the mixed integer programming model was established.According to the characteristics of the problem and model,an improved grey wolf algorithm was designed to solve the problem.The algorithm used double-layer integer coding to generate the initial solution,and improved the original grey wolf algorithm by nonlinear convergence factor and dynamic proportional weight.By combining the reverse learning mechanism and the population updating mechanism,the search solution space was further expanded,and the double threshold stop criterion was added to reduce the number of redundant iterations.Compared with LINGO solver,the results of 5~49 different scale examples showed that the model was correct and the algorithm was effective.Finally,the algorithm was used to solve the initial corridor allocation problem,and the results were compared with other algorithms,which further demonstrated the advantage of the proposed algorithm.

Key words: material loading and unloading points, asymmetric material flow, corridor allocation problem, mixed integer programming model, grey wolf algorithm

摘要： 针对制造和服务系统中假定物料装卸点重合以及设施间物流量对称的不足,结合实际生产布局对过道布置问题进行拓展,以最小化物流成本为目标,提出考虑物料装卸点及非对称流量的过道布置问题,并建立混合整数规划模型。根据问题与模型的特征,设计一种改进灰狼算法进行求解,该算法采用双层整数编码生成初始解,通过将收敛因子非线性化、比例权重动态化对原始灰狼算法进行改进。通过融合反向学习机制和种群更新机制进一步扩大搜索解空间,并添加双阈值停止准则降低多余的迭代次数。将该算法与LINGO求解器对5~49不同规模算例的计算结果进行比较,证明了模型的正确性以及算法的有效性。最后,运用该算法对初始过道布置问题进行求解,并与其他算法的求解结果进行比对,进一步证明了所提算法的优越性。

关键词: 物料装卸点, 非对称流量, 过道布置问题, 混合整数规划模型, 灰狼算法