摘要
为了有效地解决病态的约束优化问题,提出了一种模拟生物免疫系统自我调节功能的人工免疫算法,介绍了算法的基本步骤,构造了几种人工免疫算子,分析了算法的收敛性.人工免疫算法继承了遗传算法“优胜劣汰”的自我淘汰机制,但新抗体的产生方法比遗传算法中新个体的产生方法灵活得多.在进行抗体选择时若能确保当时的最优抗体可以进入下一代抗体群,则人工免疫算法是全局收敛的.100个城市TSP问题的仿真实例显示人工免疫算法比遗传算法具有更强的全局搜索能力和收敛速度.
In order to solve ill-posed constrained optimization efficiently, an artificial immune algorithm (AIA) imitating biological immune systems and its approach were presented; several artificial immune operators producing new antibodies were constituted, and the global convergence of the algorithm was studied. AIA succeeds to the mechanism of the survival of the fittest, but the methods producing new antibodies are more varied than ones producing new individuals in genetic algorithm (GA). If the best of all antibodies is made sure into next antibody colony as selecting antibodies, AIA is global convergent. A simulation text of 100-city TSP was done with AIA, and its computational process is compared with that of genetic algorithm. The results show that AIA is more efficient than GA in global optimization and computational process of AIA is faster than that of GA.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2004年第2期153-157,共5页
Control Theory & Applications
基金
国家自然科学基金项目(60272051)
教育部高校博士点专项科研基金项目(20030533014).
关键词
人工免疫算法
人工免疫系统
人工免疫算子
全局收敛性
artificial immune algorithm
artificial immune system
artificial immune operator
global convergence
