科研之友
微信
新浪微博
Facebook
分享链接
摘要
One of the fundamental problems in cellular manufacturing is grouping products with similar features into families and associated machines into cells. The objective is to maximize grouping efficacy, which indicates the within-cell machine utilization and the inter-cell movement. In this paper, a novel hybrid approach combining ordinal optimization (OO) and iterated local search (ILS) is presented to solve it in a short time. The hybrid algorithm takes ordinal optimization as the main framework, while ILS is embedded in the framework as a sub-procedure. In each iteration of the algorithm, according to OO strategy, r best solutions are accepted as the initial solutions for the embedded ILS in turn. From each initial solution, H solutions are generated and totally rH "good enough" solutions are obtained. For the ILS algorithm, a very-large scale neighborhood, cyclic transfer neighborhood is adopted, the characteristic of which is several products moving simultaneously in a cyclical manner. A reinforcement kick strategy is also proposed for the ILS algorithm, in which the products and machines are regrouped according to the current grouping relationships between products and machines. Computational experience on a set of group technology problems available in the literature shows the efficiency of the new hybrid algorithm. The results obtained by the hybrid algorithm are comparable to those obtained by other known algorithm in the literature for the problem.
