摘要
该文针对传统的遗传算法(GA)难以解决的早熟和局部收敛问题,分析了传统的GA编码策略、选择策略、交叉变异策略和交叉变异概率选择等环节存在的不足,提出一种实数编码、多种算子互相补充和交叉变异概率自适应选择的改进算法。用～个非常复杂的数学函数对新算法进行了测试,结果表明改进算法较之传统GA有效地提高了全局寻优能力。在此基础上将这种改进算法应用于热工过程辨识进行仿真研究,结果表明该方法是有效的,具有一定的应用价值,并且文中所提出的算法和策略具有一般性,很容易运用于其它优化问题。
Aiming at the problem that the traditional genetic algorithm(GA) is difficult to deal with premature and local convergence, this paper analyses the shortage of the traditional GA's primary strategy, including coding strategy, selecting strategy, crossover and mutation strategy, and selection strategy of crossover and mutation probability. An improved algorithm is proposed that adopts real coding, multi-operators cooperation and adaptive selection of crossover and mutation probability. New algorithm is tested with a complex mathematics function. The experimental results show that the improved method has better ability to converge to the global optimum than the traditional GA. The improved algorithm is applied to simulation research for thermal process identification. The results show the approach is valid. Moreover the algorithm and strategy discussed in the paper has general sense and can be applied to many other problems.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2004年第2期210-214,共5页
Proceedings of the CSEE
