Mixed-integer linear programming model for gasoline blending and distribution scheduling

Oil refineries are increasingly concerned with improving the planning of their operations and optimizing not only single production units but the whole production enterprise. However, the modeling of the overall refinery operation from the crude oil arrival to the distribution of oil products gives rise to intractable mathematical models. Thus, decomposition methodologies have long been recognized as the major avenue to overcome this computational burden. Following the spatial decomposition which distinguishes the receiving, producing, and delivery end of the refinery, the problem of gasoline blending and distribution is addressed in this paper. The problem involves the optimal operation of gasoline blending, the transfer to productstock tanks, and the delivering schedule to satisfy all of the orders. An efficient mixed-integer linear programming formulation is developed based on continuous representation of the time domain. The assumption of constant recipes is used for the blending stage. The formulation is used to address realistic case studies where feasible solutions are obtained in very reasonable computational time.