Novel memetic algorithms for flexible manufacturing systems

Flexible manufacturing systems (FMS) are systems composed of multiple heterogeneous machines and to obtain optimal schedules or solutions to FMS problems is a complex task. Evolutionary algorithms have been a popular approach to finding schedules for FMS. These algorithms, while effective, are dependent on the quality of initial populations and may not converge completely to a global optimum. This paper presents two novel memetic algorithms that combine adaptive Genetic Algorithm (AGA) with simulated annealing (SA) and local search (LS). These search techniques are used to initialize the chromosome population, enhance convergence, and refine the final schedule in GA. The resulting memetic algorithms are compared against each other and against traditional techniques (GA, SA and LS). Experimental results reveal that these memetic techniques have effectively produce improved solutions over conventional methods, often with faster convergence.