Development of multi-start simulated annealing for redesigning warehouse network problem.

Implementation of the Simulated Annealing (SA) technique is used to address a redesign warehouse network location problem. This is one of the attempts to enable the solution to escape from the local optimum. An investigation on a few elements of this meta-heuristic will be carried out in this work, The proposed SA implementation will be evaluated using the test problems from the literature. SA is one population-based, so to improve the method so it can work as well as a population-based method, multi-start techniques are discussed in this paper. The study is carried out on three types of network data: data type 1 – uniform, data type 2 – cluster, and data type 3 – compact. It shows excellent results after this method is applied. The improved average cost and very low standard deviation suggest that this method is very consistent in finding the best solution to the study problem. [ABSTRACT FROM AUTHOR]