检索结果-南通市图书馆

H∞ contr.l for.uncer.ain Mar.ovian jump.systems with mode-dep.ndent mixed delays

在线全文

学校读者我要写书评

暂无评论

p.ogr.ss in Natur.l science:Mater.als Inter.ational 2008年第3期18卷 309-314页

作者： Yingchun Wang, Huaguang Zhang school of information science and {3., northeastern univ.r.ity, shenyang 110004, china school of information science and {3. Northeastern University Shenyang 110004 China

We study the p.oblem of H∞contr.l for.a class of Mar.ovian jump.systems with nor.-bounded p.r.meter.uncer.ainties and mode- dep.ndent mixed delays including discr.te delays and distr.buted delays in this p.p.r. Our.aim is to p.esent a new delay-dep.ndent con- tr.l ap.r.ach such that the r.sulting closed-loop.system is r.bust mean-squar. (MS) exp.nentially stable and satisfies a p.escr.bed H∞p.r.or.ance level, ir.esp.ctive of the p.r.meter.uncer.ainties. Such delay-dep.ndent ap.r.ach does not r.quir. system tr.nsfor.ation or.fr.e-weighting matr.x. A numer.cal examp.e shows that the r.sults ar. less conser.ative and mor. effective.

关键词： H1 contr.l Mar.ovian jump Mixed delays Delay-dep.ndent cr.ter.on

Stability analysis of cellular.neur.l networ.s with time-var.ing delay

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Jour.al of Systems engineering and Electr.nics 2009年第2期20卷 266-273页

作者： Wang Xingang1,4, Zhang Dongmei2 & Liu Jun3.1. Coll. of information engineering, Zhejiang univ. of Technology, Hangzhou 3.003., p. r. {3.2. Coll. of science, Zhejiang univ. of Technology, Hangzhou 3.003., p. r. {3.3. Coll. of science, Beihua univ., Jilin 13.000, p. r. {3.4. school of Comp.ter.engineering and science, Shanghai univ., Shanghai 200072, p. r. china Coll. of information engineering Zhejiang Univ. of Technology Hangzhou 310032 P. R. China school of Comp.ter.engineering and science Shanghai Univ. Shanghai 200072 P. R. China Coll. of science Zhejiang Univ. of Technology Hangzhou 310032 P. R. China Coll. of science Beihua Univ. Jilin 132000 P. R. China

The global asymp.otic stability of cellular.neur.l networ.s with delays is investigated. Thr.e kinds of time delays have been consider.d. New delay-dep.ndent stability cr.ter.a ar. p.op.sed and ar. for.ulated as the feasibility of some linear.matr.x inequalities, which can be checked easily by r.sor.ing to the r.cently develop.d inter.or.p.int algor.thms. Based on the Finsler.Lemma, it is theor.tically p.oved that the p.op.sed stability cr.ter.a ar. less conser.ative than some existing r.sults.

关键词： cellular.neur.l networ.s time-var.ing delay integr.l inequality

Delay-p.obability-distr.bution-dep.ndent r.bust stability analysis for.stochastic neur.l networ.s with time-var.ing delay

在线全文

学校读者我要写书评

暂无评论

p.ogr.ss in Natur.l science:Mater.als Inter.ational 2009年第10期19卷 1333-1340页

作者： Jie Fu,Huaguang Zhang,Tiedong Ma school of information science and {3.,northeastern univ.r.ity,Box 13.,shenyang 110004,china school of information science and {3. Northeastern University Box 134 Shenyang 110004 China

The delay-p.obability-distr.bution-dep.ndent r.bust stability p.oblem for.a class of uncer.ain stochastic neur.l networ.s(SNNs) with time-var.ing delay is *** information of p.obability distr.bution of the time delay is consider.d and tr.nsfor.ed into p.r.meter.matr.ces of the tr.nsfer.ed SNNs *** on the Lyap.nov-Kr.sovskii functional and stochastic analysis ap.r.ach,a delay-p.obability-distr.bution-dep.ndent sufficient condition is obtained in the linear.matr.x inequality(LMI) for.at such that delayed SNNs ar. r.bustly globally asymp.otically stable in the mean-squar. sense for.all admissible *** imp.r.ant featur. of the r.sults is that the stability conditions ar. dep.ndent on the p.obability distr.bution of the delay and up.er.bound of the delay der.vative,and the up.er.bound is allowed to be gr.ater.than or.equal to ***,numer.cal examp.es ar. given to illustr.te the effectiveness and less conser.ativeness of the p.op.sed method.

关键词： Delay-p.obability-distr.bution-dep.ndent Stochastic neur.l networ.s Time-var.ing delay Linear.matr.x inequality

Numer.cal op.imization method for.HJI equations der.ved fr.m r.bust r.ceding hor.zon contr.l schemes and contr.ller.design

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science china.information sciences) 2012年第1期55卷 214-227页

作者： SONG ChongHui,BIAN ChunYuan,ZHANG Xie&SHI ChengLong school of information science and {3.,northeastern univ.r.ity,shenyang 110004,china school of information science and {3. Northeastern University Shenyang China

This p.p.r.addr.sses how to numer.cally solve the Hamilton-Jacobin-Isaac(HJI)equations der.ved fr.m the r.bust r.ceding hor.zon contr.l *** develop.d numer.cal method,the finite dier.nce scheme with sigmoidal tr.nsfor.ation,is a stable and conver.ent algor.thm for.HJI equations.A boundar. value iter.tion p.ocedur. is develop.d to incr.ase the calculation accur.cy with less time *** obtained value function can be ap.lied to the r.bust r.ceding hor.zon contr.ller.design of some kind of uncer.ain nonlinear.*** the contr.ller.design,the finite time hor.zon is extended into the infinite time hor.zon and the contr.ller.can be imp.emented in r.al *** can avoid the on-line r.p.ated op.imization and the dep.ndence on the feasibility of the initial state which ar. encounter.d in the tr.ditional r.bust r.ceding hor.zon contr.l schemes.

关键词： dynamic p.ogr.mming uncer.ain nonlinear.systems numer.cal method r.bust r.ceding hor.zon contr.l dier.ntial game

New adap.ive quasi-sliding mode contr.l for.nonlinear.discr.te-time systems

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Jour.al of Systems engineering and Electr.nics 2008年第1期19卷 154-160页

作者： Wang Weihong1,2 & Hou Zhongsheng3.1. school of Tr.?c and Tr.nsp.r.ation, Beijing Jiaotong univ., Beijing 100044, p. r. china.2. Dep.. of Automation, Taiyuan univ. of science & Technology, Taiyuan 03.024, p. r. china.3. Advanced Contr.l Systems Lab, school of Electr.nics and information engineering, Beijing Jiaotong univ., Beijing 100044, p. r. china school of Tr.ffic and Tr.nsp.r.ation Beijing Jiaotong Univ. Beijing 100044 P. R. China Dep.. of Automation Taiyuan Univ. of Science & Technology Taiyuan 030024 P. R. China Advanced Contr.l Systems Lab School of Electronics and Information Engineering Beijing Jiaotong Univ. Beijing 100044 P. R. China

A new adap.ive quasi-sliding mode contr.l algor.thm is develop.d for.a class of nonlinear.discr.te-time systems, which is esp.cially useful for.nonlinear.systems with vaguely known dynamics. This design is model-fr.e, and is based dir.ctly on p.eudo-p.r.ial-der.vatives der.ved on-line fr.m the inp.t and outp.t information of the system using an imp.oved r.cur.ive p.ojection typ. of identification algor.thm. The theor.tical analysis and simulation r.sults show that the adap.ive quasi-sliding mode contr.l system is stable and conver.ent.

关键词： quasi-sliding mode contr.l adap.ive contr.l model-fr.e contr.l p.eudo-p.r.ial-der.vative

Adap.ive synchr.nization of neur.l networ.s with differ.nt attr.ctor.

在线全文

学校读者我要写书评

暂无评论

p.ogr.ss in Natur.l science:Mater.als Inter.ational 2007年第6期17卷 687-695页

作者： Zhang Huaguang1, Guan Huanxin1 and Wang Zhanshan1,2(1. school of information science and {3., northeastern univ.r.ity, shenyang 110004, china.2. school of information science and {3., shenyang Ligong univ.r.ity, shenyang 110168, china. school of information science and {3. Northeastern University Shenyang 110004 China school of information science and {3. Shenyang Ligong University Shenyang 110168 China

This p.p.r.aims to p.esent an adap.ive contr.l scheme for.the synchr.nization of two classes of uncer.ain neur.l networ.s with differ.nt attr.ctor.. A new sufficient condition for.the global synchr.nization of two kinds of neur.l networ.s with differ.nt attr.ctor. is der.ved. The p.op.sed contr.l method is efficient and easy to be imp.emented. Numer.cal simulation is used to show the effectiveness of the obtained r.sult.

关键词： neur.l networ.s attr.ctor adap.ive synchr.nization linear.matr.x inequality.

Stability of discr.te Hop.ield neur.l networ.s with delay

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Jour.al of Systems engineering and Electr.nics 2005年第4期16卷 937-940页

作者： Ma r.nnian 1,2 , Lei Shep.ng3.& Liu Naigong41. Telecommunication engineering Inst., Air.For.e engineering univ., Xi’an 710071, p. r. {3.2. Key Lab of information sciences and engineering, Dalian univ., Dalian 111662, p. r. {3.3. school of Humanity Law and Economics, Nor.hwester. p.lytechnical univ., Xi’an 710072, p. r. {3.4. science Inst., Air.For.e engineering univ., Xi’an 710051, p. r. china Telecommunication engineering Inst. Air Force Engineering Univ. XT an 710071 P. R. China Key Lab of information sciences and engineering Dalian Univ. Dalian 111662 P. R. China school of Humanity Law and Economics Northwestern Polytechnical Univ. XT an 710072 P. R. China science Inst. Air Force Engineering Univ. XT an 710051 P. R. China

Discr.te Hop.ield neur.l networ. with delay is an extension of discr.te Hop.ield neur.l networ.. As it is well known, the stability of neur.l networ.s is not only the most basic and imp.r.ant p.oblem but also foundation of the networ.＇s ap.lications. The stability of discr.te HJop.ield neur.l networ.s with delay is mainly investigated by using Lyap.nov function. The sufficient conditions for.the networ.s with delay conver.ing towar.s a limit cycle of length 4 ar. obtained. Also, some sufficient cr.ter.a ar. given to ensur. the networ.s having neither.a stable state nor.a limit cycle with length 2. The obtained r.sults her. gener.lize the p.evious r.sults on stability of discr.te Hop.ield neur.l networ. with delay and without delay.

关键词： discr.te Hop.ield neur.l networ. with delay stability limit cycle.

Selection Model of Op.ional p.r.s in Sup.lier. Involved in New p.oduct Develop.ent

在线全文

学校读者我要写书评

暂无评论

武汉理工大学学报 2006年第S1期28卷 156-161页

作者： MU Lifeng,TANG Jiafu,p.N Zhendong,HAN Yi (school of information science & {3.,northeastern univ.r.ity,shenyang 110004,china. school of information science & {3. Northeastern UniversityShenyang 110004China

Quality function dep.oyment (QFD),which can tr.nslate efficiently customer.r.quir.ment (Cr.) into technical attr.butes (TAs),is widely used in new p.oduct develop.ent (Np.) as an efficient customer.dr.ven ap.r.ach. After.lots of QFD ap.lication is ap.lied in manufactur.ng fields,it was p.oved that without consider.ng sup.lier. involved,it is difficult to get p.ecise and satisfying r.sults in each p.ocess of QFD,p.r.icular.y p.r. dep.oyment. The p.p.r.will addr.ss the issues of a model of op.ional p.r.s selection in sup.lier. involved in p.r. dep.oyment p.ocess. Fur.her.it has been ap.lied to an Er.0}system r.constr.ction and up.ating p.oblem as an illustr.tion.

关键词： new p.oduct develop.ent sup.lier. involved quality function dep.oyment p.r. dep.oyment customer r.qu

Comp.r.son on neur.l networ.s and sup.or. vector.machines in sup.lier.' selection

在线全文

学校读者我要写书评

暂无评论

Jour.al of Systems engineering and Electr.nics 2008年第2期19卷 316-320页

作者： Hu Guosheng1,2 & Zhang Guohong3.1. school of Management, Guangdong Vocational Coll. of science and Technology, Guangzhou 510640, p. r. china.2. school of Comp.ter.and information, Anqing Teacher.Coll., Anqing 246011, p. r. china.3. Adult Education Coll., Guangzhou univ., Guangzhou 510405, p. r. china school of Management Guangdong Vocational Coll. of Science and Technology Guangzhou 510640 P. R. China school of Comp.ter.and information Anqing Teacher Coll. Anqing 246011 P. R. China Adult Education Coll. Guangzhou Univ. Guangzhou 510405 P. R. China

Sup.lier.＇ selection in sup.ly chain management （SCM） has attr.cted consider.ble r.sear.h inter.sts in r.cent year.. r.cent liter.tur.s show that neur.l networ.s achieve better.p.r.or.ance than tr.ditional statistical methods. However. neur.l networ.s have inher.nt dr.wbacks, such as local op.imization solution, lack gener.lization, and uncontr.lled conver.ence. A r.latively new machine lear.ing technique, sup.or. vector.machine （SVM）, which over.omes the dr.wbacks of neur.l networ.s, is intr.duced to p.ovide a model with better.exp.anator. p.wer.to select ideal sup.lier.p.r.ner.. Meanwhile, in p.actice, the sup.lier.＇ samp.es ar. ver. insufficient. SVMs ar. adap.ive to deal with small samp.es＇ tr.ining and testing. The p.ediction accur.cies for.Bp.N and SVM methods ar. comp.r.d to choose the ap.r.ciating sup.lier.. The actual examp.es illustr.te that SVM methods ar. sup.r.or.to Bp.N.

关键词： sup.lier.selection sup.ly chain management logistics sup.or. vector.machine Bp.N.