简介:Apotentialreductionalgorithmisproposedforoptimizationofaconvexfunctionsubjecttolinearconstraints.Ateachstepofthealgorithm,asystemoflinearequationsissolvedtogetasearchdirectionandtheArmijo'sruleisusedtodetermineastepsize.Itisprovedthatthealgorithmisgloballyconvergent.Computationalresultsarereported.
简介:ItisprovedthatalinearlyrecursivesequenceofnindiceaoverfieldF(n≥1)isautorntaticallyaproductofnlioearlyrecurplvesequenceaof1-lndexoverFbythetheoryofHopfalgebras.Bytheway,thecorrespondencebetweenthesetoflinearlyrecursivesequenoesof1-indexandF[X]^0isgeneralisedtothecaseofn-index.