Dynamic scheduling of blocking flow-shop based on multi-population aco algorithm

Abstract

Drawing on the coevolution between populations, this paper proposes a dynamic multi-population ant colony optimization (ACO) algorithm to solve the blocking flow-shop scheduling problem (FSP). In our algorithm, the ant colony is divided into an elite population, multiple search populations, and a mutation population. In the initial stage, only the elite population and the search populations participate in optimization. After a certain number of iterations, a mutation population is dynamically generated from the worst solution in each search population and that in the elite population. The mutation population is reinitialized before entering the optimization process. The mutated population can jump out of the original search space for another search. Finally, the superiority of our algorithm in solving blocking FSP was proved through comparative simulations.