摘要
经典的蚁群算法模仿蚂蚁觅食, 释放信息素, 形成正反馈互激励机制, 提高了全局寻优效率, 但它只适用于离散问题. 将解空间划分为小区域, 用以承载信息素, 设置全局与局部蚂蚁, 引入遗传算法的种群和操作方式, 以Powell寻优算子和最优解保留策略改造蚂蚁的智能活动与互激励机制, 构建为杂交蚁群系统 (hybridant colony system, HACS), 可用于求解连续优化问题. 实例测试表明, HACS具有良好的全局寻优能力和稳定性, 将HACS应用于2 氯苯酚在超临界水中氧化反应动力学参数的估算, 获得了满意的结果.
A drawback of standard ant colony algorithm is not suitable for solving continuous optimization problems. A hybrid ant colony system (HACS), which was combined with the genetic algorithm and the Powell method, was proposed for solving continuous optimization problems. HACS, which included two kinds of ants: global ant and local ant, was built. The function of global ants was to explore the solution space, and that of local ants was to exploit the best solution. These two kinds of ants cooperated with each other in finishing the optimization task. The Powell method joined in local search, which was in favor of obtaining the accurate global optimal solution. The testing results indicated that the HACS showed better ability in finding global optimal solution than SGA. HACS was used to estimate the kinetic parameters of 2-chlorophenol oxidation in supercritical water.
出处
《化工学报》
EI
CAS
CSCD
北大核心
2005年第3期487-491,共5页
CIESC Journal
基金
国家自然科学基金项目 (20276063).~~
