Efficient chromosome encoding and problem-specific mutation methods for the flexible bay facility layout problem

Abstract

Two chromosome encoding methods are compared for finding solutions to the nondeterministic polynomial-time hard flexible bay facilities layout problem via genetic algorithm (GA). Both methods capitalize on the random key GA approach to produce chromosomes that are viable for any combination of allele values. In addition, the effect of four problem-specific mutation methods is assessed for one of the encoding methods. The novel mutation methods are shown to have a substantial effect on performance. Optimal GA parameter settings for the problem-specific mutation methods are found empirically