Optimization of the multi-objective transhipment problem with hybrid fuzzy evolutionary algorithm

We consider a system with n stores or stocking locations, which may differ in their demand parameters. Stores have to order products periodically to fill up their demands. Stores review their inventory periodically and procure the product. If a store runs out of stock or has surplus inventories transshipment takes place here to reconcile. Transshipments between stocking locations are used to minimize total costs and maximize service levels. The objectives to be satisfied in our problem are the minimization of total costs, lead times, and quality deficiency while maximizing fill rates. The reason that the problem becomes more complex and harder to solve is the nature of conflicting objectives and also the noisy decision variables. Evolutionary algorithms are able to propose efficient solutions in both cases, therefore it is convenient to use an evolutionary algorithm method. We propose a hybrid fuzzy evolutionary algorithm based on the strength Pareto evolutionary algorithm 2 (SPEA2) method to solve the problem. The proposed algorithm employs a fuzzy inference system to the selection operator, in an attempt to improve the algorithm performance. Empirical results show that the proposed hybrid fuzzy evolutionary algorithm for the multi-objective transshipment problem produces appropriate solutions and can be efficiently used to solve multi-objective problems. This approach provides Pareto optimal solutions which give the chance to choose from a set of best solutions. Furthermore, this method can also be applied to alike multi- objective problems to obtain Pareto optimal solutions.