Stochastic programming with pricing decision for refinery planning

Abstract

Planning and scheduling of refinery involve the management of crude oil purchasing and products production schedule for reaching the maximum limit of economic margin. The development of modeling and optimization for refinery planning and scheduling to response the dynamic environment is underway. Another important aspect for planning in a highly dynamic evnironment is price-decision making. This study proposed a planning model with pricing decision for refinery operation under uncertainty based on two-stage stochastic approach. The starting point was the development of the planning model. The relationship between product prices and demand was modelled and forecasted. Then it was integrated into a mathematical planning model in order to simultaneously determine the prices and the optimal schedule. The objective function was based on optimizing the profit by maximizing sale and minimizing cost. The consideration of the uncertaity and financial risk management were introduced. Then the imput/output interfaces of the model were created to make the model more user-friendly. The model was tested/run by applying the data from the Bangchak Petroleum Public Company Limited. The optimization results show that the planning model with pricing decision suggests a lower expected profit but it is more reliable when compared to the non-pricing one and the stochastic model results can predict a higher profit than that of the deterministic model.