Assessing hypermutation operators of a clonal selection algorithm for the unequal area facility layout problem

Abstract

A mutation operator is critical for the performance of a clonal selection algorithm (CSA) since it diversifies the search directions and avoids early convergence to local optima. This article introduces a CSA approach for the unequal area facility layout problem (UAFLP) with flexible bay structure. A new encoding, the use of mutation types with different combinations, and different static and dynamic mutation application strategies are also proposed. In addition, a guideline in parameter optimization of the CSA is provided. An experimental study is performed on five cases of the UAFLP. It is concluded that the hypermutation types studied in this article, especially the inverse mutation followed by pairwise mutation, can be used to obtain good results within short computation times.