Some new dynamic economic lot sizing models

We introduce a dynamic economic lot sizing problem which takes into account both the production setup cost and the designed and yet stretchable production capacity. Demands and the control on production activities drive the evolution of the production system. We give a pseudo-polynomial algorithm for the problem when future demands are known. When there is uncertainty in future demands and there is no holding of inventory, we propose a two-layered solution process which makes the demand acceptance decision first and decides the level of production next. We also construct a mixed model which combines the features of both deterministic and stochastic models. The model not only utilizes currently-known information but also anticipates for the future. Our simulation study demonstrates the advantage of the mixed model.