求解线性约束问题的微粒群优化算法.pdfVIP

10 4 ( ) V ol. 10 No. 4 2010 12 JOURNA L OF NAN JING NORMA L U NIV ERS ITY ( ENG INEER ING AN D T ECHNOLOGY ED IT ION) Dec, 2010 求解线性约束问题的微粒群优化算法 陈战平1, 2 ( 1. , 210046; 2. , 210097) [ ] , . , , , . , , . . [ ] , , [ ] TP 3016 [] A [] 2010) Particle Swarm Algorithm for L inear Constrained Optim ization Problem C hen Z hanp ing1, 2 ( 1. S chool of Compu te Science and T echnology, Nanjing N om al Un ive s ity, Nanj ing 210046, Ch in a; 2. Jiangsu Resea ch Cen te of In fo m ation S ecu ity and P ivacy T echnology, Nanjing 210097, Ch ina) Abstract: The sho tage in slow conve gence ate and low conv e gence p ec ision ex istw hen the pa ticle swa m a lgo ithm is d i ectly used to so lve the const a ined optmi ization p oblem. In th is pape , w e a e conce ned w ith an new pa ticle swa m algo ithm, w hich can be used to so lve the linea const a ined p ob lem s. In ou me thod, the const ained optmi iza tion p oblem is fi st t ans lated into a nonconst a ined optmi ization one by int oducing the L ag ang e mu ltip lie s, and then by us ing the Lag ange dua lity p inc ip le, the L ag ange mu ltip lie s and op tmi ization pa am ete s a e sepa ated, which w ill be optmi ized espective ly by using the pa ticle swa m a lgo ithm. M o eove , in o de to m ake the pa tic le sw a m a lgo ithm conv e ge to the g loba l optmi ization so lution, an mi p oved pa ticle swa m a lgo ithm w ith muta tion is p opo sed. F ina lly, a design exam ple of a lowpass F IR filte show s tha t ou m ethod is be tte than the pa ticle sw a m al