Stochastic optimization of two-stage multi-item inventory system with hybrid genetic algorithm

This paper considers a two-stage, multi-item inventory system with stochastic demand. First we propose two types of exact stochastic optimization models to minimize the long-run average system cost under installation and echelon (r, nQ) policy. Second we provide an effective hybrid genetic algorithm (HGA) based on the property of the optimization problem. In the proposed HGA, a heuristic search technique, based on the tradeoff between inventory cost and setup cost, is introduced. The long-run average cost of each solution in the model is estimated by Monte Carlo method. At last, computation tests indicate that when variance of stochastic demand increases, echelon policy outperforms installation policy and the proposed heuristic search technique greatly enhances the search capacity of HGA.