Artificial bee colony algorithm and its application to permutation flow job scheduling problem

Flow shop scheduling problem (FSSP) is probably one of the most well known industrial engineering problem of the past 50 years. Although there are exact solutions for one or two machines, the problem becomes NP-hard for three or more machines. This challenging environment attracts many researchers to offer more efficient solution algorithms. Artificial Bee Colony (ABC) algorithm is a relatively new swarm intelligence-based algorithm proposed by Karaboğa (2005), inspired by the intelligent foraging behaviors of honeybee swarms. Primarily used for solving numerical optimization problems, it is now used widely in different engineering problem types. In this study, ABC algorithm is applied to permutation flow shop scheduling problem. Performance of the algorithm is tested by using Carlier, Reeves and Taillard test instances and compared against different metaheuristic-based methods. For the Carlier and Reeves instances, the proposed ABC algorithm produced better results than most of the known metaheuristic algorithms in the literature. A new best solution is gained for the Reeves23 test problem. Some competitive results are found for the Taillard test problems against different metaheuristic methods used for comparing. In some instances, ABC algorithm performed better than the other methods, whereas in some instances its solutions are slightly worse than the others.