Multiperiod mixed integer programming models and techniques for new petroleum refinery construction

Gasoline producers all over the world are now preparing for new, stricter governmental regulations on impurity levels of refined petroleum products. Due to their environmental, health and other implications, certain ingredients will have to appear in smaller proportions than they do today. In order to be able to produce gasoline, and other refined oil products, satisfying the future constraints, refineries may need to be expanded by the construction of new refining plants. We model the question of whether to invest in new plants as a multiperiod, mixed integer programming problem. The contribution of this paper is a discussion of some modelling issues and some solution techniques for an investment problem of direct relevance to the petroleum industry.