?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=The+Effect+of+Offspring+Population+Size+on+NSGA-II%3A+A+Preliminary+Study&rft.creator=Hort%2C+M&rft.creator=Sarro%2C+F&rft.description=Non-Dominated+Sorting+Genetic+Algorithm+(NSGA-II)+is+one+of+the%0D%0Amost+popular+Multi-Objective+Evolutionary+Algorithms+(MOEA)%0D%0Aand+has+been+applied+to+a+large+range+of+problems.%0D%0APrevious+studies+have+shown+that+parameter+tuning+can+improve%0D%0ANSGA-II+performance.+However%2C+the+tuning+of+the+offspring+population+size%2C+which+guides+the+exploration-exploitation+trade-off+in%0D%0ANSGA-II%2C+has+been+overlooked+so+far.+Previous+work+has+generally%0D%0Aused+the+population+size+as+the+default+offspring+population+size+for%0D%0ANSGA-II.%0D%0AWe+therefore+investigate+the+impact+of+offspring+population+size%0D%0Aon+the+performance+of+NSGA-II.+We+carry+out+an+empirical+study+by%0D%0Acomparing+the+effectiveness+of+three+configurations+vs.+the+default%0D%0ANSGA-II+configuration+on+six+optimization+problems+based+on+four%0D%0APareto+front+quality+indicators+and+statistical+tests.%0D%0AOur+findings+show+that+the+performance+of+NSGA-II+can+be+improved+by+reducing+the+offspring+population+size+and+in+turn+increasing+the+number+of+generations.+This+leads+to+similar+or+statistically%0D%0Asignificant+better+results+than+those+obtained+by+using+the+default%0D%0ANSGA-II+configuration+in+92%25+of+the+experiments+performed.&rft.subject=Genetic+algorithms%2C+multi-objective+optimization%2C+NSGA-II%2C+offspring+population&rft.publisher=Association+for+Computing+Machinery+(ACM)&rft.date=2021-07-14&rft.type=Proceedings+paper&rft.publisher=Genetic+and+Evolutionary+Computation+Conference&rft.language=eng&rft.source=+++++In%3A++Proceedings+of+the+2021+Genetic+and+Evolutionary+Computation+Conference+(GECCO+2021).++++Association+for+Computing+Machinery+(ACM)%3A+New+York%2C+NY%2C+USA.+(2021)++++(In+press).++&rft.format=text&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10129908%2F1%2FNSGA-II_Offspring.pdf&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10129908%2F&rft.rights=open