SINGLE ROW FACILITY LAYOUT PROBLEM USING AN IMPERIALIST COMPETITIVE ALGORITHM

In this paper, a novel metaheuristic algorithm named imperialist competitive algorithm (ICA) is employed to solve the single row facility layout problem (SRFLP). SRFLP is known as NP-hard and involves arranging a number of facilities on a straight line to minimize transportation costs among the facilities. ICA starts with an initial population of countries (solutions) which are classified as Imperialist and Colony. Each colony in the initial population is assigned to an imperialist respectively. An imperialist and its corresponding colonies construct an Empire. ICA proceeds through three main components, namely, assimilation inside each empire, imperialistic competition among empires and elimination of powerless empires. ICA converges to local optima when there is only one empire left. Details of applying ICA to SRFLP are elaborated and the quality of ICA is evaluated on a number of SRFLP instances (30 small-size instances and 20 large-size instances) from the literature. For small-size instances with known optimal solutions, ICA was able to find all the optimal solutions. In addition, for large-size instances with no proven optimal solutions, ICA succeeded in improving 8 previously known best solutions.