An Improved Differential Evolution Algorithm for Operating Optimization of a Distillation Unit

This paper studies a challenging optimization problem of operating variables for a distillation unit (DU) widely used in oil refineries. The problem aims to optimize the operating variables when a new environment (e.g., the feedstock properties change) arises. We first establish the operation optimization model of the DU. Subsequently, we propose an improved differential evolution (IDE) algorithm that meets the problem characteristics to solve this optimization problem. In the IDE algorithm, a mutation operator is used to obtain a mutant individual (also called mutant vector), and an adaptive scheme is proposed to update the scaling factor used in the mutation strategy to balance the exploitation and exploration capabilities of the algorithm. We introduce an effective strategy to correct infeasible operating variables in the mutant vector to improve the optimization efficiency of the algorithm. Experiments show the superiority of the IDE algorithm compared to other algorithms.