Multi-guided particle swarm optimization : a multi-objective particle swarm optimizer

Abstract

English: An exploratory analysis in low-dimensional objective space of the vector evaluated particle swarm optimization (VEPSO) algorithm is presented. A novel visualization technique is presented and applied to perform the exploratory analysis. The exploratory analysis together with a quantitative analysis revealed that the VEPSO algorithm continues to explore without exploiting the well-performing areas of the search space. A detailed investigation into the influence that the choice of archive implementation has on the performance of the VEPSO algorithm is presented. Both the Pareto-optimal front (POF) solution diversity and convergence towards the true POF is considered during the investigation. Attainment surfaces are investigated for their suitability in efficiently comparing two multi-objective optimization (MOO) algorithms. A new measure to objectively compare algorithms in multi-dimensional objective space, based on attainment surfaces, is presented. This measure, referred to as the porcupine measure, adapts the attainment surface measure by using a statistical test along with weighted intersection lines. Loosely based on the VEPSO algorithm, the multi-guided particle swarm optimization (MGPSO) algorithm is presented and evaluated. The results indicate that the MGPSO algorithm overcomes the weaknesses of the VEPSO algorithm and also outperforms a number of state of the art MOO algorithms on at least two benchmark test sets.

isiZulu: Uhlaziyo oluhlolayo olusendaweni yomgomo yesilinganisobungako esiphansi sendlela yokwenza yobungako obunobukhulu nenkombandlela obuhloliwe benhlayiya yoluquqaba yokuthola isixazululo esingcono kakhulu luyethulwa. Isu eliyisibonakaliso senoveli liyethulwa futhi lisetshenzisiwe ukwenza uhlaziyo oluhlolayo. Uhlaziyo oluhlolayo kanye nohlaziyo oluphathelene nokubala luveze ukuthi indlela yokwenza yobungako obunobukhulu nenkombandlela obuhloliwe benhlayiya yoluquqaba yokuthola isixazululo esingcono kakhulu iyaqhubeka nokuhlola ngaphandle kokuxhaphaza izindawo ezisebenza kahle zendawo yophenyo. Uphenyo olugcwele mayelana nomthelela ukukhetha kokuqaliswa kwengobo yomlando okunakho ekusebenzeni kwendlela yokwenza yobungako obunobukhulu nenkombandlela obuhloliwe benhlayiya yoluquqaba yokuthola isixazululo esingcono kakhulu luyethulwa. Kokubili ukungafani nokuhlangana kwesisombululo esingabuswa esinye esisombululweni esingabuswa ngesinye esiliqiniso kuyabhekwa ngesikhathi sophenyo. Izindawo ezingaphandle zokuthola ziphenyelwa ukufaneleka kwazo ekuqhathaniseni ngempumelelo izindlela zokwenza ezimbili zokuthola isixazululo esingcono kakhulu okunemigomo eminingi. Isikali esisha sokuqhathanisa ngokungakhethi izindlela zokwenza endaweni yomgomo yezilinganisobungako eziningi, ngokubhekisa ezindaweni ezingaphandle zokuthola, siyethulwa. Lesi sikali, esibizwa ngokuthi yisikali se-porcupine, sijwayeza isikali sendawo engaphandle yokuthola ngokusebenzisa ithuluzi lezibalo kanye nolayini abaphambanayo abanesisindo. Ngokubhekisa okungatheni endleleni yokwenza yobungako obunobukhulu nenkombandlela obuhloliwe benhlayiya yoluquqaba yokuthola isixazululo esingcono kakhulu, indlela yokwenza yenhlayiya yoluquqaba yokuthola isixazululo esingcono kakhulu enemihlahlandlela eminingi iyethulwa futhi iyahlolisiswa. Imiphumela ikhombisa ukuthi indlela yokwenza yenhlayiya yoluquqaba yokuthola isixazululo esingcono kakhulu enemihlahlandlela eminingi idlula ubuthakathaka bendlela yokwenza yobungako obunobukhulu nenkombandlela obuhloliwe benhlayiya yoluquqaba yokuthola isixazululo esingcono kakhulu iphinde futhi idlule kude le inqwaba yezindlela zokwenza zokuthola isixazululo esingcono kakhulu okunemigomo eminingi eziseqophelweni eliphezulu okungenani emasethini ebhentshimakhi amabili okuhlola.

Afrikaans: 'n Verkennende ontleding in lae-dimensionele objektiewe ruimte van die vektor-geëvalueerde deeltjieswermoptimisering (VEPSO)-algoritme word aangebied. n Nuwe visualiseringstegniek word aangebied en toegepas om die verkennende ontleding uit te voer. Die verkennende ontleding, tesame met 'n kwantitatiewe ontleding, het aan die lig gebring dat die VEPSO-algoritme voortgaan om te verken sonder om die goed presterende areas van die soekruimte te ontgin. 'n Gedetailleerde ondersoek na die invloed wat die keuse van argiefimplementering op die werkverrigting van die VEPSO-algoritme het, word aangebied. Beide die Pareto-optimale front (POF) oplossingsdiversiteit en konvergensie na die ware POF word tydens die ondersoek oorweeg. Bereikingsoppervlaktes word ondersoek vir hul geskiktheid om twee multi-objektieweoptimisering- (MOO) algoritmes doeltreffend te vergelyk. n Nuwe maatstaf om algoritmes in multidimensionele objektiewe ruimte objektief te vergelyk, gebaseer op bereikingsoppervlaktes, word aangebied. Hierdie maatstaf, waarna verwys word as die ystervarkmaatstaf, pas die bereikingsoppervlakmaat aan deur 'n statistiese toets saam met geweegde snylyne te gebruik. Losweg gebaseer op die VEPSO-algoritme, word die multigeleide deeltjieswermoptimisering (MGPSO)-algoritme aangebied en geëvalueer. Die resultate dui daarop dat die MGPSO-algoritme die swakhede van die VEPSO-algoritme oorkom en ook 'n aantal moderne MOO-algoritmes op ten minste twee maatstaftoetsstelle oortref.

Sepedi: Tshekatsheko ya go nyakišiša sekgobeng sa maikemišetšo a mahlakore a fase a algoritheme ya go dira gore go be le dikarolwana tša sehlopha sa go hlahlobja (VEPSO) yeo e hlahlobjago ya vector e tšweletšwa. Thekniki ye mpsha ya pono e tšweletšwa le go dirišwa go phethagatša tshekatsheko ya go utolla. Tshekatsheko ya go nyakišiša mmogo le tshekatsheko ya dipalo e utollotše gore algoritheme ya VEPSO e tšwela pele go hlahloba ntle le go šomiša mafelo ao a šomago gabotse sekgobeng sa go utolla. Go tšweletšwa ka nyakišišo ka botlalo ya khuetšo yeo kgetho ya phethagatšo ya polokelong e nago le yona go tshepedišo ya algoritheme ya VEPSO. Bobedi, tharollo ya "Pareto-optimal front" (POF)ye fearologanego le kopanyo ya dilo go ya go POF ya nnete di ile tša elwa hloko nakong ya nyakišišo. Bokagodimo bja phihlelelo bo a nyakišišwa go lekola go swanelega ga tšona bja go bapetša bokgoni bja dialgoritheme tše pedi tša maikemišetšo a mantši tša go dira gore dilo di šome gabotse (MOO) . Tekanyo ye mpsha ya go bapetša dialgorithemo ka nepo sekgobeng sa maikemišetšo a mahlakore a mantši, yeo e theilwego bokagodimong bja phihlelelo, e tšweleditšwe maikemišetšo a mantši. Tekanyo ye, yeo e bitšwago tekanyo ya porcupine, e tlwaetša phihlelelo ya tekanyo ya bokagodimo ka go šomiša teko ya dipalopalo gotee le mela ya magahlanong ye e nago le boima. Ka motheo wo lokologilego woo o theilwego godimo ga algoritheme ya VEPSO, algoritheme ya go dira gore dihlopha tša dikarolwana tše dintši tše di hlahlago (MPGSO) e a tšweletšwa le go hlahlobja. Dipoelo di laetša gore algorithm ya MGPSO e fenya mafokodi a algorithm ya VEPSO gomme gape e feta palo ya dialgorithemo tša MOO tša maemo a godimo ka bonnyane disete tše pedi tša diteko tša tekanyetšo.