A two-step heuristic algorithm for layout design of unequal-sized facilities with input/output points

The facility layout problem (FLP) is to arrange facilities within a given area so as to minimise the total material-handling cost. An efficient arrangement of facilities plays an important part in the achievement of high productivity in a manufacturing system. This paper focuses on the layout design of unequal-sized facilities with fixed shapes and input/output (I/O) points. A mixed-integer programming (MIP) model is developed to obtain optimal solutions. Since the MIP model is not applicable to large-sized problems, a two-step heuristic algorithm is developed to solve the FLPs. In the first step, a layout solution with moderate quality is generated by using an interconnected zone algorithm and a simulated annealing (SA) algorithm. The zone algorithm is a newly developed layout construction technique which places facilities successively within designated zones according to a placing sequence. The zone concept is introduced to reduce the solution space such that the computational effort is decreased, while SA is used to search for the placing sequences. The second step improves that solution further using the reduced MIP by employing the relative position information obtained from the first step. Computational experiments indicate that the algorithm in the first step can yield on average a 1.048% improvement on the best solutions obtained by the existing algorithms. The second step can further improve the solutions obtained in the first step by 0.484% on average. Compared to the commercial software VIP-PLANOPT 2006 (Engineering Optimization Software 2005 Engineering Optimization Software. 2005. PLANOPT User’s Manual, Florida: Deltona. [Google Scholar]), the proposed two-step algorithm improves the solutions by 9.113% on average.