Optimization of Pooling Problems for Two Objectives Using the ε-Constraint Method

Abstract

Pooling problems are optimization problems of importance in petroleum refineries. They contain bilinear terms, which lead to nonconvex optimization problems, each with multiple minima. Hence, a global optimization method is required to find the optimal solutions to pooling problems. The solutions to pooling problems for single objectives have been studied using many deterministic global optimization algorithms. However, there has been no attempt to solve industrially important pooling problems for multiple objectives. Hence, in this chapter, pooling problems are optimized for two objectives using the ε-constraint method along with a recent stochastic global optimization algorithm, namely, integrated differential evolution. Further, a new formulation that does not involve equality constraints for pooling problems is described and used. Using this formulation, many pooling problems from the literature are optimized for two objectives, and the results demonstrate the potential of multi-objective optimization for finding tradeoff solutions for pooling problems. © 2013 John Wiley & Sons, Ltd.