The modular facility layout consists of grouping machines into modules according to subsequences of operations common to a set of parts. In the method proposed by Huang (2003), facility layout modules are generated by cluster analysis. This paper introduces an alternative way to generate layout modules through a clustering genetic algorithm (CGA). The CGA allows the user to specify the number of modules he wants (K-grouping problem) and to optimize the number of modules (automatic clustering problem). An important feature of the genetic algorithm is its flexibility, which enables the user to interact with the problem-solving method by presenting him with a choice of chromosome codes, measures of similarity to compare sequences of operations, and methods for generating new populations. To this end, two crossover operators and three mutation operators were developed. In the example presented here, the layout modules generated by the CGA and the plant's final physical layout were compared with those of Huang (2003). This comparison revealed a significant reduction in the total distance covered by the set of parts, demonstrating the effectiveness of the clustering genetic algorithm. Thus, the results indicated that the clustering genetic algorithm is an alternative for generating layout modules in the design of modular facility layouts.
Modular facility layout; Genetic algorithms; Clustering; Manufacturing
