简介：Inthispaper,wepresentanalgorithmforsolvingaclassoftransportationproblemsandshowthatitisgloballyconvergent.Thepropsedalgorithmissimpletoimplementandeasytoparallelize.Thenumericalresultsindicatethattheproposedalgorithmiseffectiveforthisclassofproblems.

简介：Thepaperpresentsaneuralnetworkforsolvingaclassofquadraticprogrammingproblems.Theneuralnetworkiscompletelystabletoexactsolutionsandtherearenoparameterstoset.Moreover,noanaloguemultipliersanddividersarerequired,incontrasttoexistingneuralnetwork[3]whichneedsplentyofanaloguemultipliers.

简介：Thebasicprobleminoptimizingcommunicationnetworksistoassignapropercircuitforeachorigindestinationpairinnetworkssoastominimizetheaveragenetworkdelay,andthenetworkoptimalrouteselectionmodelisamulti-constrained0-1nonlinearprogrammingproblem,Inthispaper,anewstochasticoptimizationalgorthm,ImmuneAlgorithm,isappliedtosolvetheoptimizationproblemincommunicationnetworks,AndthebackbonenetworkvBNSischosentoillustratethetechniqueofevaluatingdelayinavirtualnetowrk.Atlast,IAiscomparedwiththeoptimizationmethodincommunicationnetworksbasedonGeneticAlgorithm,andtheresultshowsthatIAisbetterthanGAinglobaloptimumfinding.

简介：Theapplicationofcellularneuralnetworks(CNN)forsolvingpartialdifferentialequations(PDEs)isinvestigatedinthispaper.TwokindsofthePDEs,theheat-conductionequationandPoisson'sdquation,areconsideredtobetypicalexamples.TheycanbecomputedinrealtimebyusingtheCNN,whiletheCNN'shardwareisimplementedbytheintegratedOP-AMP.Theexperimentalresultsshowthatthehardwareperformenceisinagreementwiththatgivenbythecomputersimulation.Therefore,theCNNisanewpowerfultoolforsolvingPDEs.

简介：Thekeyideabehindculturalalgorithmistoexplicitlyacquireproblem-solvingknowledgefromtheevolvingpopulationandinreturnapplythatknowledgetoguidethesearch.Inthisarticle,culturalalgorithm-simulatedannealingisproposedtosolvetheroutingproblemofmobileagent.Theoptimalindividualisacceptedtoimprovethebeliefspace’sevolutionofculturalalgorithmsbysimulatedannealing.Thestepsizeinsearchisusedassituationalknowledgetoguidethesearchofoptimalsolutioninthepopulationspace.Becauseofthisfeature,thesearchtimeisreduced.Experimentalresultsshowthatthealgorithmproposedinthisarticlecanensurethequalityofoptimalsolutions,andalsohasbetterconvergencespeed.Theoperationefficiencyofthesystemisconsiderablyimproved.

简介：反碰撞的技术是在收音机频率鉴定(RFID)的一个关键点系统。避免数据碰撞，有二条途径：阿罗哈基于算法，二进制树(BT)基于算法。然而，这些不能完全解决碰撞问题，特别当标签数量大，标签标志长时。在这篇文章，我们在场一个多分支询问树(MBQT)协议基于平衡的不完全的块设计(BIBD)，代码，和使用16位向量作为RFID的询问前缀符号源于BIBD阅读器。与常规反碰撞算法相比，理论上的分析和模拟证明建议协议改进鉴定效率。