基于神經(jīng)網(wǎng)絡(luò)與模糊評判的正交設(shè)計系統(tǒng)研制與開發(fā).pdf

1、東北大學(xué)碩士學(xué)位論文基于神經(jīng)網(wǎng)絡(luò)與模糊評判的正交設(shè)計系統(tǒng)研制與開發(fā)姓名：馮鐵申請學(xué)位級別：碩士專業(yè)：機械設(shè)計及理論指導(dǎo)教師：隋天中2003.2.1查!!壟堂堡主蘭堡壘墨一——————』旦墜墨壘望ADevelopmentandStudyontheSystemaboutOrthogonalDesignbasedonFuzzyAnalysisMethodandArtificialNeural_NetworkAbstractTheOrthogo

2、nalDesign，whichplaysallextremelyimportantroleontechnologyandqualityinnovation，isanadvancedexperimentalprojectandoptimizationtechnologyandusedinnearlyeveryfieldsAlthoughOrthogonalDesignusedtobepopularizedtocertainextantin

3、ourland，OrthogonalDesignitselfhastremendousshortcomingsBecauseofthese，itisnecessarytocombineotherrelativelyadvancedtechnologyinordertoinprovethesimpleOrthogonalDesignmethodFuzzyintegratedevaluation，withtheoryofthefuzzyma

4、th，isabletodrawarelativelyexplicitconclusionbyjudgingofsomecertainfuzzyinformationAswearejudgingthenon—numericinformation，wemuchmoreneedfuzzyanalysisBecausetheremerelyabsolutelydistinctionsbetweeneverytargetandlevel，weha

5、vetoneedsomemethodsuchasFuzzyEvaluationTherearemuchmoreapplicationsofthemethodinthepracticeTherefore，itiswisetocombinethetwomethodsorganicallyexeaingsuperioritiesofthetwosu街cientlyInthiswayWCcanefficientlyovercometheshor

6、tagecausedbysimplytheOrthogonalDesignInthiscondition，theOrthogonalDesignCombinedfuzzyevaluationhascomeoutAfterhavingobtainedsomefeasibleprojects，wecanusefuzzymethodtodealwiththedatacausedbytheOrthogonalDesignmethod，obtai

7、ningaSmuchinformationaspossibleThen，wecallsufficientlyusetheinformationtoobtaintheinteractionofparametershandledandthebestcompositivetechnicalcondition，havingtransformedthequalitydefinitiontothemassevaluationandcombinedt

8、hem，namelyhavingcombinedsubjectiveandobjectiveArtificialNeural—Networkhadbeenbeingdevelopedsince1980s，whichcallsimulateselfhistology—learningabilityofbiologies’nerveandismadeupofmanycollaterallysimplecalculatingunitsThen