进化规划-蚁群优化算法的构建并用于化工过程操作优化

Construction of EP-ACO and its application in operation optimization of chemical process

经典蚁群优化(ACO)算法搜优效率高,但只适用于求解组合优化等离散问题.以搜索最优食物源为目标,并引入进化规划(EP)简洁的进化机制,用以改造ACO,使之适于连续问题.又将蚁群分工为全局和局部蚂蚁,分别引领个体进行全局探索式和局部挖掘式寻优,并在各个体上释放信息素,供蚁群共享,由此继承了ACO正反馈、互激励的优点,并在优进策略的支持下,构建为EPACO算法.经复杂测试函数的优化检验,显示出EPACO适于连续问题,且全局搜优效率高,对高维问题适应性强.将EPACO应用于二甲苯异构化装置的操作优化,取得了良好的效果,与其他方法相比,优越性明显.

Ant colony optimizat combinational optimization prob on （ACO） has high optimizing efficiency, but can only be applied to ems. For adapting ACO to continuous optimization problems, the concise evolution mechanism of evolution program （EP） was introduced to reconstruct ACO, in which the objective was to search optimal food source other than the best sequence. The ant colony was divided into global ants and local ants, which guided the individuals to perform global exploratory optimization and local excavating optimization respectively. Ants released pheromone on the individuals, and the pheromone was shared by all ants, which inherited the collective autocatalytic behaviour characterised by positive feedback mechanism of ACO. Under the support of eugenic strategy, the EP ACO algorithm was constructed. The experimentations on optimization of complex functions showed that EP-ACO could be well fit for solving continuous optimization problems with high global optimization efficiency and showed good adaptability to high dimension problems. Finally, EP-ACO was successfully applied to the operation optimization of the equipment of xylene isomerization. The results were better than the referenced methods.