Post-optimality Analysis of Production Levels in a Petrochemical Complex

This study investigates the effect of uncertainty or variation associated with some model parameters on the production rates in a petrochemical complex. A modified approach to post-optimality analysis is used to assess and evaluate such effects. Frequent and small variations in parameters, like product prices, or constraints, like demand and supply, can seriously affect the optimum solution. To overcome this problem, various fuzzy and stochastic programming techniques have been proposed in the literature to determine robust solutions. However, application of post-optimality analysis in the petrochemical industry has received less attention. In this paper, a modified method for post-optimality (stability) analysis is applied to a petrochemical complex model formulated as a linear programming problem. The proposed approach differs from stochastic programming in that it can provide the decision-maker with simple and useful information that enables identification of sensitive parameters and constraints that need better estimation and monitoring. The results present the stability limits within which the assigned production rates remain optimum and sensitivity information like the Lagrange multiplier is still valid. By applying these results, the profit of the petrochemical complex can be increased by between 3.4 and 6.3 % by adjusting some model constraints.