Comparison between Three Metaheuristic Algorithms for Minimizing Cycle Time in Cyclic Hybrid Flow Shop Scheduling with Learning Effect

Jobs scheduling in industries with cyclic procedure on machines, such as perishable products (food industries) or products with a limited lifetime (chemicals, radio actives, etc), is very important. Due to time limitation or competition with other companies, these industries try to minimize thecycle time of jobs processing. Since most productive environments of the industries are cyclic hybrid flow shop and operator’s learning effect is obvious in speed of productions, the aim of this study is to minimize cycle time of each machine with learning effect by consequence of jobs. After proposing a mathematical model and since the cyclic hybrid flow shop environment is NP-hard, three metaheuristics, i.e., genetic algorithm, simulated annealing algorithm and population based simulated annealing algorithm, have been proposed for solving this problem. Results show that on average, population based simulated annealing algorithm due to its population-based structure has a better performance in comparison to other algorithms.