用于多目标优化的蚁群算法的构建及其应用

Ant colony optimization approach to the solution of multi-objective optimization problems

从蚁群的生物学行为出发，将成群募集和海量募集两种机制融入蚁群算法，并针对多目标优化的特性，综合考虑解的被支配度和分散度，抽提出一种启发式规则，用以评价食物源的优劣，进而构建多目标连续蚁群优化算法（MO-CACO）。通过两个多目标典型函数的优化测试，验证了MO-CACO具有较强的多目标全局寻优能力，且稳健性良好，所求得的最优解集的多目标值能均匀地逼近Pareto最优前沿的各部分。将MO-CACO用于二甲苯异构化装置的操作优化，取得了满意的结果，MO-CACO可为化工过程多目标决策提供支持。

In this article, a multi-objective continuous ant colony optimization （MO-CACO） for solving multi-objective optimization problem is proposed, which is based on the biology behavior of ant colony foraging and the heuristic rules for appraising the quality of food source. The mechanisms of group recruitment and mass recruitment, which were used to guide the ant colony to search the best solution in the feasible region, were embedded to the ant colony system. The proposed algorithm was applied to two benchmark functions to illustrate its effectiveness. The results demonstrated that MO-CACO had better performance for achieving global optimal and the obtained multi-objective values can unifonnly approximate every part of Pareto-optimal front. Further, it was successfully applied to the optimization of the equipment of xylene isomerization, and the satisfying results were obtained, which can be used for further decision-making.