Multi-objective optimization of manufacturing cell design

Abstract

Whereas the single-objective cell-formation problem has been studied extensively during the past decades, research on the multi-objective version of the problem has been relatively limited, despite the fact that it represents a more realistic modelling of the manufacturing environment. This article introduces multi-objective GP-SLCA, an evolutionary computation methodology for the solution of the multi-objective cell-formation problem. GP-SLCA is a hybrid algorithm, comprising of GP-SLCA, a genetic programming algorithm for the solution of single-objective cell-formation problems, and NSGA-II, a standard evolutionary multi-objective optimization technique. The proposed methodology is capable of providing the decision maker with a range of non-dominated solutions instead of a single compromise solution, which is usually produced as an outcome of alternative multi-objective optimization techniques. The application of multi-objective GP-SLCA is illustrated on a large-sized test problem taken from the literature.