太原理工大学硕士研究生学位论文
THERESEARCH OFSTRUCTURALLEARND呵G
METHODBASEDONQUOTIENTSPACE
ABSTRACT
Asan emerging researchsub一fieldofartificial intelligence,ganular
computing,认七。se户1105叩冲 isto加PlemenitheProblemsolving atdi月飞rent
leve lsofgranularity,a1mstoestab1ishmuchmore genera1modelrefiecting the
processof human Problem solving .G.口ule,a clumP of Points
(objects) 山aWn togetherbyindistingu1shability, similarity, Pro劝mity or
允nctionality,isthe Primitivonotion ofgranularcompu石ng.
Quotlent sP‘etheory ofProblem solving isgeneralizedfrom 七刀0
directions,n田nefythestructUreoftheuniverseofdiscours eandgranu1ation
criteria.Inthelight ofgranu1arcomPuting quotient sPace the0ry ofProb1em
solvingexP1oitst0Pologytodescribethe structureofthe un1verseofdiscourSe,
叨dutil让eseqp1valenc ere1ation toimPletnent granu1ation,曲drelles on the
natoralmapping toreal泳 the translationsamongdifferentleve isofgranulariy.
Asamethed ofProblem solving,quotlent sPacethe o理,based on
subs栩Lntlal theory,cons ideringtheproblem from differentasPects and
mu lti七1erarchyinthe Pr0cessofPr0blem solving , isakindofPowerfoltoolin
thatitcan deereasethedifficultyoftheProblem andreducethe comPu枉妓ional
cost.Un ify1ngthequantitativeanaly sisandthequa1itative ana lysisby utilizing
ITI
评论0
最新资源