Abstract
In this paper, we present a new heuristics for the facility layout problem (FLP). It is known that the FLP can be solved to optimality via mixed integer programming (MIP). The largest problem that has been solved to optimality to date, however, contains only eleven departments, and thus, is not applicable to typical problems where there are many more departments. In this paper, we developed a new force-directed model that enables us to develop good relative locations of departments very quickly by exploiting the special structure of the problem. We used the chaos characteristics of the force-directed model used in the algorithm so that they do not become stuck as a result of the local optima. We use this relative position to formulate the FLP as linear programming. In the computational experiments, we tested our proposed method against several well-known benchmark data sets. Our algorithm shows better or equal cost performance in most of them, and also finishes the computations much faster.
