DocumentCode :
2451988
Title :
An Adaptive Genetic Algorithm Based on Multi-population Parallel Evolutionary and Variable Population Size
Author :
Chen, Jianxin ; Liu, Qing ; Huang, Junqin ; Hou, Yun
Author_Institution :
Key Lab. of Special Area Highway Eng. of Minist. Educ., Chang´´an Univ., Xi´´an, China
Volume :
1
fYear :
2010
fDate :
16-17 Dec. 2010
Firstpage :
258
Lastpage :
262
Abstract :
The paper deals with the performance of genetic algorithm according to the analysis of control parameters, the evolution between different sub-populations, the interaction on best individual and, the expansion on the search interval, are analyzed, and an adaptive genetic algorithm is presented, which is multi-population parallel evolutionary and variable population size. It is proved by comparative experiments that the speed of convergence and the precision of the new algorithm are considerably improved, which avoid the premature convergence phenomenon of single-population evolutionary algorithm, and maintain the evolutionary stability of the best individuals, so it effectively makes up the, shortcomings of single-population and constant parameters, which don´t overcome the premature phenomenon universally and so on. The results show that it is better than the traditional one both robustness and effectiveness of the algorithm. Therefore, the algorithm in practice has a broad application prospects.
Keywords :
adaptive control; convergence; genetic algorithms; nonlinear control systems; robust control; adaptive genetic algorithm; control parameters; convergence speed; evolutionary stability; multipopulation parallel evolutionary; robustness; search interval; variable population size; Accuracy; Algorithm design and analysis; Convergence; Diversity reception; Evolutionary computation; Libraries; Stability analysis; adaptive; genetic algorithm; parallel evolutionary; variable population size;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Systems (GCIS), 2010 Second WRI Global Congress on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-9247-3
Type :
conf
DOI :
10.1109/GCIS.2010.102
Filename :
5708756
Link To Document :
بازگشت