基于进化算法的鲁棒优化方法与其在模拟电路设计中的应用-研究

《基于进化算法的鲁棒优化方法与其在模拟电路设计中的应用-研究》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、AbstractABSTARCTTherearealotofoptimizationproblemsinengineeringpracticeandscientificresearch．Thetraditionaloptimizationproblemsfocusontheglobaloptimalsolutions．Theactualperformanceofthesolutionswhichareverysensitivetointerferenceisoftennotgood．Tllisisbecaus

2、ealotofnoiseaffectstheactualperformanceoftheorysolutions，forexample，theparameterdrift,temperaturechangesandotherinterference．Inconsiderationofinterferenceintheactualenvironment，therobustoptimizationfocusesonboththeperformanceandrobustnessofsolutions．Inreali

3、tytheoptimizationproblemsareallrobustoptimizationproblemsbecauseoftheubiquitousinterference，SOthereisawiderangeofpracticalsignificancetostudytherobustoptimization．nlepaperstudiesontherobustoptimizationandtheirapplicationinanalogcircuitdesign．Themaincontribu

4、tionincludes：1．ExperimentalstudyoftherobustoptimalsolutionsiSdoneinthethesis．Therobustnessevaluationmethodsbasedontheaverageandtheworstcasearesummarized．Wedesignsomeexperimentstotesttheeffectofevolutionaryalgorithmforrobustoptimizationproblem，alsoabouttheke

5、yfactorsinthisproblem，forexample，theinterferencefieldsize，samplingfrequency，samplingmethod．2．Tosolvetherobustoptimizationproblembasedontheaverageevaluation，weproposeanimprovedalgorithmbasedonthebootstrapmethods．TheimprovedalgorithmCanapproximateeffectiveobj

6、ectivefunctionmoreaccuratelywithoutincreasingthesamplingfrequency．Theadvantagesofthealgorithmareverifiedontheexperiment．Wealsostudyinfluenceofthekeyfactorsontheperformanceofthealgorithm．3．Tosolvetherobustoptimizationproblembasedontheworstcase，wetransfertheo

7、riginalproblemintominimaxproblem，andusethecoevolutionaryalgorithmtofindtherobustoptimalsolutions．Theadvantagesofthealgorithmareverifiedontheexperiment．Theimportantfactorsarealsostudied，forexample，therealfitnessandthefitnessdeviation．4．Analogcircuitsareextre

8、melysensitivetointerference，andevolutionarydesignofanalogcircuitsistime-consumingandcostly．Tosolvethis，thenovelrobustoptimizationalgorithmbasedonbootstrapmethodisappliedtotheparameteroptimizationofanal