简介：Thispaperelaboratestheconceptandmodelofprotocolmigrationinnetworkinterconnection.Migrationstrategiesandprinciplesarediscussedandseveralcasesarestudiedindetailwhichshowthebasicprocedureandtechniquesusedinprotocolmigration.

简介：Wehavedevelopedafuzzyneuralexpertsystemthathastheprecisionandlearningabilityofaneuralnetwork.Knowledgeisacquiredfromdomainexpertsasfuzzyrulesandmembershipfunctions.Then,theyareconvertedintoaneuralnetworkwhichimplementsfuzzyinferencewithoutrulematching.Theneuralnetworkisappliedtoproblem-solvingandlearnsfromthedataobtainedduringoperationtoenhancetheaccuracy.Thelearningabilityoftheneuralnetworkmakesiteasytomodifythemembershipfunctionsdefinedbydomainexperts.Also,bymodifyingtheweightsofneuralnetworksadaptively,theproblemofbeliefpropagationinconventionalexpertsystemscanbesolvedeasily.Convertingtheneuralnetworkbackintofuzzyrulesandmembershipfunctionshelpsexplaintheinnerrepresentationandoperationoftheneuralnetwork.