简介：Thispaperpresentsaconcurrentobject-orientedmodelinglanguagebasedonPetrinets:OMNets,whichhelpsgreatlytoavoidtheinheritanceanomalyproblemappearedinconcurrentOOlanguages.OMNetsseparatesthefunctionalpartandthesynchronizationpartofobjectsandusesPetrinetstodescribethesynchronizationpart.Bothpartsarereusablethroughinheritance.

简介：给一个n维的格子L和某目标向量，这份报纸为近似最靠近的向量问题学习算法(由使用近似最短的独立向量问题神谕的CVP)(SIVP)。更精确，如果在目标向量之间的距离和格子不比c\大({\scriptscriptstyle\frac{c}{\gamman}}{\uplambda}_1\left(\mathrm{L}\right)\)为任意的大却有限的经常的c>0，我们给使随机化并且发现最靠近的向量的确定的多项式时间算法，当以前的减小仅仅为\被知道时({\scriptscriptstyle\frac{c}{2\gamman}}{\uplambda}_1\left(\mathrm{L}\right)\)。而且如果在目标向量和格子之间的距离比关于用SIVP神谕和Babais最近的飞机算法的n(L),的某数量大，我们能解决\(\mathrm{CVP}\upgamma\sqrt{n}\)在确定的多项式时间。特殊，如果近似因素(1，2)在SIVP神谕，我们为CVP获得一个更好的减小因素。

简介：InthispaperwestudythesolutionofSATproblemsformulatedasdiscretedecisionanddiscreteconstrainedoptimizationproblems.Constrainedformulationsarebetterthantraditionalunconstrainedformulationsbecauseviolatedconstraintsmayprovideadditionalforcestoleadasearchtowardsasatisfiableassignment.Wesummarizethetheoryofextendedsaddlepointsinpenaltyformulationsforsolvingdiscreteconstrainedoptimizationproblemsandtheassociateddiscretepenaltymethod(DPM).Wethenexaminevariousformulationsoftheobjectivefunction,choicesofneighborhoodinDPM,strategiesforupdatingpenalties,andheuristicsforavoidingtraps.ExperimentalevaluationsonhardbenchmarkinstancespinpointthattrapscontributesignificantlytotheinefficiencyofDPMandforceatrajectorytorepeatedlyvisitthesamesetofornearbypointsintheoriginalvariablespace.Toaddressthisissue,weproposeandstudytwotrap-avoidancestrategies.Thefirststrategyaddsextrapenaltiesonunsatisfiedclausesinsideatrap,leadingtoverylargepenaltiesforunsatisfiedclausesthataretrappedmoreoftenandmakingtheseclausesmorelikelytobesatisfiedinthefuture.Thesecondstrategystoresinformationonpointsvisitedbefore,whetherinsidetrapsornot,andavoidsvisitingpointsthatareclosetopointsvisitedbefore.Itcanbeimplementedbymodifyingthepenaltyfunctioninsuchawaythat,ifatrajectorygetsclosetopointsvisitedbefore,anextrapenaltywilltakeeffectandforcethetrajectorytoanewregion.Itspecializestothefirststrategybecausetrapsarespecialcasesofpointsvisitedbefore.Finally,weshowexperimentalresultsonevaluatingbenchmarksintheDIMACSandSATLIBarchivesandcompareourresultswithexistingresultsonGSAT,WalkSAT,LSDL,andGrasp.TheresultsdemonstratethatDPMwithtrapavoidanceisrobustaswellaseffectiveforsolvinghardSATproblems.