Decomposition andHierarchical Process forFuzzyCognitive Mapsof Complex Systems
Zhang Gui-yun · 2006
ofconcepts andtheir relations. Concept (denoted byverInorder toreduce thescale andcomplexity, thede- tex) canrepresent activity, cause, effect, intent, emotion, composition methods offuzzy cognitive mapsforknowl-proneandtrends ofsystem. Itreflects theattribute, peredgerepresentation andreasoning arepresented. Parti-formance, quality andstate. Relations ofconcepts repretionthevertices oforiginal cognitive mapintosome sent therelationship between concepts andeffect relation groups, andthen construct quotient cognitive mapbased(denoted byanarcwitharrow, thedirection ofarrow onthese groups. Thus, theanalysis oftheoriginal FCM represents thedirection ofrelationship andeffect), its canbereduced tothat ofquotient FCM,which providesstrength isrepresented bynumber value[']. global information oforiginal FCM,andsomesubFCMs Asanewtechnology method ofknowledge reprewhichkeepanoriginal topological structure andinfer-sentation andreasoning, thecognitive mapofAxelord is enceandprovide local information oforiginal FCM.It amodeltorepresent causal relation inrepresentation and extremely reduces theoriginal sizeandcomplexity, inference system, andprovides aneffective soft computmeanwhile, itispossible tostudy system atdiferent ingtool tosupport self-adapted behavior based ontranslices andenhance theenrichment andflexibility ofre- scendental knowledge. Cognitive mapcanrepresent dysearch between concepts. Someknowledge representa- namiccausal system withfeedback, whichisdifficult to tions andreasoning ofcomplex system usethemethod of berepresented bytreestructure, Bayesnetwork and hierarchicalforf uzzy cognitive maps. Therefore, ithigh-Markov. Recently, thetheory ofFCM hasbeenfound lights thehierarchical causal relationship andinclusion manyapplications inpolitics, economics, medicines, relationship between concepts, andembodies thespacemilitary, social relation andinformation system, network causal relation between concepts. administration, exception analysis, multi-agent system, industrial control andvirtual reality etc[2'3].