一类具有精英学习能力的增强型人工免疫网络优化算法 - 控制理论与应用.pdf

26 3 Vol. 26 No. 3 2009 3 Control Theory Applications Mar. 2009 :2009 , , , (, 510275) : (Enhanced aiNet–EL). , . , Enhanced aiNet–ELaiNetEaiNet. , PID. , Enhanced aiNet-EL 4. : ; ; ; ; PID : TP273 : A An enhanced artificial immune network with elitist-learning capability for optimization problems LI Zhong-hua, ZHANG Yu-nong, TAN Hong-Zhou, CHEN Zhuo-yi (School of Information Science and Technology, Sun Yat-sen University, Guangzhou Guangdong 510275, China) Abstract: This paper proposes an new enhanced artificial immune network with elitist-learning (Enhanced aiNet-EL) for optimization problems. The proposed new algorithm integrates affinity-learning with elitist-learning and its three immune operators, i.e., cloning, mutation and suppressor. The simulation results on two classical benchmarks indicate that the proposed enhanced aiNet-EL optimization outperforms the traditional aiNet optimization and EaiNet optimization in both the final solution and convergence speed. In applying the proposed algorithm to an industrial PID control system, the step response shows a performance better than those under other four approaches. Key words: artificial immune network; elitist-learning; affinity-learning; particle swarm optimization; PID controller 1 (Introduction) . aiNet , , , , , . — . ( /), , . . , , , . , , . , de Castro . (aiNet),