简介：Therecentlyinventedartificialbeecolony(ABC)algorithmisanoptimizationalgorithmbasedonswarmintelligencethathasbeenusedtosolvemanykindsofnumericalfunctionoptimizationproblems.Itperformswellinmostcases,however,therestillexistsaninsufficiencyintheABCalgorithmthatignoresthefitnessofrelatedpairsofindividualsinthemechanismoffindinganeighboringfoodsource.ThispaperpresentsanimprovedABCalgorithmwithmutuallearning(MutualABC)thatadjuststheproducedcandidatefoodsourcewiththehigherfitnessbetweentwoindividualsselectedbyamutuallearningfactor.TheperformanceoftheimprovedMutualABCalgorithmistestedonasetofbenchmarkfunctionsandcomparedwiththebasicABCalgorithmandsomeclassicalversionsofimprovedABCalgorithms.TheexperimentalresultsshowthattheMutualABCalgorithmwithappropriateparametersoutperformsotherABCalgorithmsinmostexperiments.

简介：ＡｒｔｉｆｉｃｉａｌＩｎｔｅｌｌｉｇｅｎｃｅＥｍｂｅｄｄｅｄＯｂｊｅｃｔ－ＯｒｉｅｎｔｅｄＭｅｔｈｏｄｏｌｏｇｙＦｏｒＭｏｄｅｌＢａｓｅｄＤｅｃｉｓｉｏｎＳｕｐｐｏｒｔ￥ＦｅｎｇＳｈａｎ；ＴｉａｎＹｕａｎ；ＬｉＴｏｎｇ＆ＣａｉＪｕｎ（Ｉｎｓｔｉｔｕｔ...

简介：极化控制器(PC)是为极化(懦夫)的状态的不可缺少的被动部件在光纤维通讯系统的控制。以便克服存在极化控制器(PC)的劣势，新奇PC在这份报纸基于人工的birefringence被设计。从3阶段延迟的并置的PC结果控制其各个由二极化横梁splitters，二棱柱和一个piezoz-lectric变换器(PZT)组成的单位压电陶器。试验性的结果证明建议PC能把一个任意的输入懦夫变换成任何期望的输出有在低控制电压下面的高精确性的。

简介：Handwrittensignaturerecognitionispresentedbasedonananglefeaturevectorbyusingtheartificialneuralnetwork(ANN)inthisresearch.Eachsignatureimagewillberepresentedbyananglevector.ThefeaturevectorwillconstitutetheinputtotheANN.Thecollectionofsignatureimagesisdividedintotwosets.OnesetwillbeusedfortrainingtheANNinasupervisedfashion.TheothersetwhichisneverseenbytheANNwillbeusedfortesting.Aftertraining,theANNwillbetestedbyrecognizingthesignatures.Whenasignatureisclassifiedcorrectly,itisconsideredcorrectrecognition,otherwiseitisafailure.Theachievedrecognitionrateofthissystemis94%.