Vendor selection and purchasing decisions

We would like to determine a minimal set of vendors and optimal order quantities from selected vendors to maximize quality, minimize late delivery and minimize cost including business volume discounts while satisfying the component demands under capacity limits of the vendors. In reality, the demands are usually unknown. Moreover, the capacities of vendors are not constant. Our stochastic programming model treats demand and capacity as random variables. We compare our stochastic programming model with recourse with a deterministic mixed integer programming model. Numerical experiments show that capturing uncertainty in a stochastic programming model provides more robust solutions than a deterministic mixed integer programming model.