hatareaddedintoapopulation,sothattoimprovethequalityofthepopulation.thenumericalsimulationandpracticaldesignexamplesshowthatthethreenovelgeneticalgorithmshavemuchbetterconvergenceabilitiesthanthestandardgeneticalgorithm(seefan,xi,wang,inverseproblemsinengineering,2000,).
2.thegenetic-algorithm-baseddesignprinciplesandmodelsarefirstsystematicallyestablished(seefan,journalofpowerandenergy,proc.itn.mech.engrs,1998).inthegenetic-baseddesignmodels,basedontheimprovedgeneticalgorithms,thetuningandoperationpatterofthegeneticoperatorsinadesignoftheaerodynamiccascadesareexplored.someparameterizatioandtheircorreondingcodingmethodsforaerodynamiccascadesregardingtotheoperatioofgeneticalgorithmsarepresented.inthemeanwhile,intheproposeddesignmodels,incorporatinggeneticalgorithmsandartificialneuralnetworksisattemptedtosolvecascadedesignproblems.inthiscase,agenetic-algorithm-baseddesignmethodisembeddedwithafeedforwardartificialneuralnetworkthatisusedtocomputetheflowcharactersofgivebladeprofiles.astheresult,thefitnecomputationaltimecanbereduced,andfurtherthealgorithm’sevolvingepochcanbeshortened.moreover,thefeedforwardartificialneuralnetworksarefirstuseddirectlytocompleteacascadeaerodynamicdesigntask,withthegeneticalgorithmsbeingusedtotrainandevolvethenetwork.allthegenetic-baseddesignmodelsestablishedpoewidegeneralities.theycanincorporatewithanydegreecfdsolvers.theycanimplementautomaticdesigofanaerodynamiccascadeinanyrequireddesig,i.e.,aconventionaldirectdesignorinversedesign,andahybriddesi

上一页  [1] [2] [3] [4] [5] [6] [7] 下一页