Two new models and an algorithm for stochastic lot-sizing problems

The single-stage, capacitated lot-sizing problems with multiple products and random demands are considered in this paper. A chance-constrained programming model and an expectation model under stochastic environment based on real production strategies are presented for the case of deterministic capacity constrained and the case of uncertain capacity respectively. A genetic algorithm based on stochastic simulation is used to solve both of the models. The paper studies the effects of various model parameters and environmental factors on the objective costs using analysis of variance, and concludes with some advice for future research activities in the end.