基于多目标进化算法的对二甲苯氧化反应过程优化操作

Optimization of p-xylene oxidation reaction based on multi-objective evolutionary algorithm

针对单目标优化算法在求解优化问题时普遍存在易陷入局部最优的缺陷,本文提出采用基于精英选择和个体迁移的多目标优化方法来求解单目标优化问题,通过对原问题目标进行有效分解,将其分解为多个子目标,通过对多个子目标的优化来扩大搜索范围,加快算法的收敛速度。仿真结果表明,与文献中方法相比,采用多目标方法来求解单目标问题能显著提高算法收敛速度,将其应用于对二甲苯(PX)氧化反应过程的优化操作,在相同计算成本条件下,醋酸和PX燃烧损失明显下降,成本损失大幅减少。

Aiming at the limitation that the single-objective optimizing algorithm is generally easy to trap into local optimal solution,multi-objective evolutionary algorithm based on elitist selection and individual migration was proposed for single-objective optimization problem.Decomposing the original single objective into multi-objectives,it could broaden its searching extension and quicken its converging speed.Simulation results showed that the proposed method could obviously speed up its converging performance.Furthermore,the application to p-xylene oxidation reaction indicated that the proposed method could obviously reduce acetic acid and p-xylene combustion loss,greatly reduce production cost at the same numerical computing cost.