Comparison Study ofSPEA2+, SPEA2, andNSGA-II inDiesel Engine Emissions andFuelEconomyProblem

Recently, thetechnology that cancontrol NOx andSootvalues ofdiesel engines bychanging theelec- tronically controllable parameters hasbeendeveloped. However, there isatrade-off relationship between fuel economy andNOxvalues. Therefore, thediesel engines that canchange their characteristics withalong tothe driving environment should beemerged inthefuture. Fordesigning these kinds ofengines, thePareto solu- tions that canexpress thetrade-off between fuel econ- omyandNOxvalues areneeded. Inthat case, thede- rived nondominated solutions should havethediver- sity notonlyintheobjective spacebutalsointhede- sign variable space. SPEA2+isoneofmulti objective genetic algorithms andisdeveloped based onSPEA2. Thederived nondominated solutions bySPEA2+have thediversity inbothobjective space anddesign variable space. Inthis study, thediesel engines that havehigh fuel economyandsmall amountsofNOxandSootare designed bySEPA2+.Theresults arecompared with those ofSPEA2andNSGA-II. Fromthediscussions, it isfound that thesolutions ofSPEA2+havethediversity notonly intheobjective space butalso inthedesign vari- ablespace. Thesecharacteristics areverysuitable for designing diesel engines whoseparameters arechanging against thedriving environment.