欢迎来到天天文库
浏览记录
ID:36457964
大小:2.32 MB
页数:63页
时间:2019-05-10
《基于人工鱼群算法的信号MP稀疏分解》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、西南交通大学硕士学位论文基于人工鱼群算法的信号MP稀疏分解姓名:舒维杰申请学位级别:硕士专业:信号与信息处理指导教师:尹忠科20080701西南交通大学硕士研究生学位论文第1I页AbstractThesignalsparsedecompositionisemergingitsanewmethodforanalyzingandprocessingsignals,whichhasmanyexcellentcharacteristics.AndsignalsparsedecompositionbasedonMatchingPursu
2、it(MP)isusedcommonlyinsignalsparsedecomposition.Ithasbeenappliedtomanyareassuchasdatacompression,signalfeatureextraction,time—frequencyanalysisandetc..ButitisaNPdifficultproblem.Thelargecomputationalcostisthebottleneckofsparsedecomposition.Inrecentyears,swarmintelli
3、gencealgorithmhasattractedlotsofresearchesofnumerousscienceareas,whichhashigheradaptability,robustness,parallelandglobalqualityetc.,andareusedwidelyinsomeareassuchasfunctionoptimization,patternrecognitionandimageprocessingetc.Thisarticlestudiestheapplicationofanewsw
4、armintelligencealgorithm—ArtificialFish·SwarmAlgorithm(AFSA),insolvingsignalsparsedecompositionandtherelatedapplication.Thisevolutionalgorithmcanovercomelocalextremumandgetsomeglobalextremums,initialvaluescallbechosenrandomandcarryingoutthisalgorithmdoesn’tneedtokno
5、wgradientvalueoftheobjectivefunction,SOithasrobustnessforthesearchingarea.ThepaperstartswiththesignalsparsedecompositionandMatchingPursuitalgorithm.AfterwardsArtificialFish-SwarmAlgorithmisintroduced.Itpresentssystemicexpatiationandstudyabouttheprinciple,structure,a
6、stringencyandimplementsmethodsofthealgorithm.Andanimprovedartificialfish-Swarmalgorithmisproposed.Inthepreyactionofartificialfish,thepullmeansisaddedwhichcancollectmoreartificialfishneartheglobalextremum.TheImprovedalgorithmnotonlyretainstheprecisionoforiginalalgori
7、thm,butalsoadvancestheefficiencyandconvergentspeed.ThentheimprovedArtificialFish-SwarmAlgorithmisappliedtothesignalsparsedecompositionbasedonMP,andthesimulationresultsandanalysisaregiven.ItshowsthattheimprovedArtificialFish.SwarmAlgorithmCanfastsearchforapproximatel
8、yoptimalatomateachstepofMPandthequantumofcomputingisreducedalot.Inaddition,itisinevitableto西南交通大学硕士研究生学位论文第1II页●
9、II_——_——_——_————_—●———●●_
此文档下载收益归作者所有