Adaptive Range Particle Swarm Optimization to Find Multiple Optima(Mechanical Systems)
Satoshi KITAYAMA, Chie Miyakawa, Koetsu YAMAZAKI, Masao ARAKAWA · TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series C · 2009
uf En 暮i1 ］ e じrfllg． Kanaza 、 ヤ a Urliverlsily ， Kakurna -mach1 ， Kanazawrshl ， 罫 shikawa ， 920 -1］92 Japとm The maill objective callcd as the populaLion based optirnizati ⊂ ｝n techrLiques is 匸o fil ／d a global minirnum ， HQwever ， iしis somc し ilrles preferableto find i-ultiple optirna ill tl ｝eengineerillg design． The Adapdve Range ParLicle Swarm ( 〕 pt． innizutio1 〕 ( ARPSO )1． ⊂ 〕 find Tnu ］ tiple optirl ／a is proposed il コ thls paper ． The ac 膳 ． e sear ・ ch domain range of the ARPSO is only one ， however ． many actiyc search domain ranges t(浦 nd multiple opthna arc generated in the proposed appoac ｝ 1、 At nrst， the manv activc search domain ranges are generat εd by the paired particles， and the procedui ． e 〔 ｝f the ARPSO is appl ［ ed ． Then ， the paired partic ［ es find to(：al or global Millilllum separetely ． Many active search doinaiii r ' anges arc sometimes intagrated through the scarch process ， The dctail Dr ' Dc ( 三 durビ t｛， integrate the Hctive scarch dQnnain rangcs is also explained ill this paper ， 1亡 is P ｛ )ssible to fincl multiple optill 〕a with high ； ICcUTlcQcY by the proposed approach ． Th ' e validity 〔 〕f prop 〔 )sed me 亡 hod is exannined through numerical examples ．