A comparison of two chromosome representation schemes used in solving a family-based scheduling problem

Meta-heuristic algorithms have been widely used in solving scheduling problems; previous studies focused on enhancing existing algorithmic mechanisms. This study advocates a new perspective—developing new chromosome (solution) representation schemes may improve the performance of existing meta-heuristic algorithms. In the context of a scheduling problem, known as permutation manufacturing-cell flow shop (PMFS), we compare the effectiveness of two chromosome representation schemes (Sold and Snew) while they are embedded in a meta-heuristic algorithm to solve the PMFS scheduling problem. Two existing meta-heuristic algorithms, genetic algorithm (GA) and ant colony optimization (ACO), are tested. Denote a tested meta-heuristic algorithm by X_Y, where X represents an algorithmic mechanism and Y represents a chromosome representation. Experiment results indicate that GA_ Snew outperforms GA_Sold, and ACO_Snew also outperforms ACO_Sold. These findings reveal the importance of developing new chromosome representations in the application of meta-heuristic algorithms.

► We study the effect of chromosome representations on the performance of GA and ACO. ► A scheduling problem (PMFS) is used as the problem context. ► Numerical experiments reveal that GA_Snew comprehensively outperforms GA_Sold. ► Numerical experiments indicate that ACO_Snew comprehensively outperforms ACO_Sold. ► This finding highlights the value of exploring new solution representation schemes.