基于募集机制的连续蚁群系统及其应用

Recruitment mechanism based continuous ant colony optimization system and its application

针对经典蚁群算法只适用于离散优化问题的不足,从蚂蚁觅食的生物学行为出发,以寻觅最优食物源为目标,运用蚁群的海量募集和成群募集两种机制,并结合蚂蚁的厌食现象,均衡地搜索,由此构建了适用于连续问题的蚁群优化系统(MG-CACO).经典函数的测试表明,MG-CACO的全局寻优效率高,稳健性良好,尤其对高维问题的适应性强.将MG-CACO用于二甲苯异构化装置的操作条件优化,效果令人满意,其全局寻优性能和稳定性均优于其他方法.

Traditional ant colony algorithm is suitable for solving discrete optimization problem, but is rarely used for continuous optimization problem. A new ant colony optimization system called MG-CACO for solving continuous optimization was proposed based on the biological behavior of ant colony foraging. The mechanism of mass recruitment, group recruitment, and the phenomenon of anorexia were embedded into the ant colony optimization system. The performance of MG-CACO was demonstrated through two benchmark functions, the testing results indicated that MG-CACO has good stability and ability for finding the global optimal value. MG-CACO was successfully applied to the optimization of the equipment of xylene isomerization, with the obtained results being better than the reference methods.