关键词: 装配线平衡, 物料超市规划, 协同优化, 遗传算法

Abstract: With the increasing demand for multiple varieties and customization, manufacturers should design assembly lines efficiently to enhance the competitiveness of their products. Recently, manufacturers began to widely use new material supermarkets near assembly lines to ensure just in time part supply of stations. In the assembly line design stage, the assembly line balancing problem and the supermarket planning problem are directly interrelated ones. The decision taken to solve assembly line balancing problem will limit the layout of the material supermarket, and the material supermarket planning will affect the logistics efficiency and cost of the assembly line.
At present, most studies use a hierarchical approach to separate these two interrelated problems, which balance the assembly line to get the optimal number of workstations at first, and then decide the number and location of material supermarkets. The optimal solution in the first step limits the result of the material supermarket planning, and may cause an increase in total costs.In this paper, the model and algorithm for the collaborative optimization on assembly line balancing and material supermarket planning are proposed. Based on the problem description and related assumptions, an integrated mixed integer programming model is constructed with the goal of minimizing the total cost including workstation installation cost, material supermarket installation cost and transportation cost. In order to solve the large-scale problems, an improved genetic algorithm which applies a novel encoding mode and a new population initialization method is proposed.
To evaluate the computational performance of the proposed improved genetic algorithm, problems of different scales are employed in the numerical analyses. The results obtained from the proposed improved genetic algorithm are compared with results solved by Cplex and the traditional genetic algorithm. The comparison demonstrates that the proposed improved genetic algorithmhas better performance as the scale of the problem increases. It can get the optimal solutions in a short time. Moreover, the comparison result shows that the collaborative optimization approach can get lower total cost compared with the hierarchical approach.

Key words: assembly line balancing, material supermarket planning, collaborative optimization, genetic algorithm