is presented. The advantages of dynamic encoding and adjustment of genetic operators in MLGA are analyzed. And improved crossover and mutation techniques in MLGA are proposed. The correlated variables
described as points
in individuals can be operated with crossover and mutation simultaneously. These techniques ensure that all the individuals are limited in effective search space. This MLGA is applied in optimization design of power transformer successfully. A comparison with single level ordinary GA shows that MLGA reduces the cost of transformer by 1.83% and the no-load loss by 1.01% with similar computational consumption.
PEI Suping, WANG Xiaolei, CHEN Yuguo, et al.The study of improved genetic algorithm in transformer optimization design[J]. Journal of Zhongyuan Institute of Technology, 2006,17(2):52-55.
FAN Shuwei, WANG Guoliang, XIE Wei. Genetic algorithm and its application in power transformer optimization design[J].Proceeding of the CSEE, 1996,16(5):346-348.
WU Derong, LI Jingchuan, LI Qingfu, et al. Design optimization of series electrical machines using genetic algorithms [J]. Journal of Xi'an Jiaotong University, 1999, 33(2):14-18.
WU Zhongchao,YAN Xinping. A multi-parallism genetic algorithm and its application[J]. Journal of Wuhan Transportation University, 2000, 24(4): 384-387.
LI Q S, LIU D K, LEUNG A Y T, et al. A multilevel genetic algorithm for the optimum design of structural control systems [J]. International Journal for Numerical Methods in Engineering, 2002,55:817-834.
JENKINS W M. On the application of natural algorithms to structural design optimization[J]. Engineering Structures,1997, 19(4):302-308.