表1　数值计算结果

4　结　语

本文引入了武器目标分配问题的一个非线性整数规划模型,鉴于非线性整数规划的难以求解性,本文利用WTA非线性整数模型中目标函数的特点,使用线性函数来替代原问题中的非线性目标函数,近而将非线性整数规划问题转化为线性整数规划问题。并给出了初步的算例来验证算法的有效性。线性整数规划属于难以找到多项式时间解的问题,对于任意规模,计算时间会较长。但本文中给出的方法较为简单,能够在较短的时间内给出一个近似最优解,实施武器目标分配方案,满足实战的需要。参考文献:

算例规模(W×T)5×510×1015×1520×2030×30

算例序号

1212121212

最优解-702.2-700.8-1946.4-1893.5-2934.6-2147.4-4341.1-4736.8-5815.7-6045.7

计算时间(s)2.282.282.292.282.402.282.762.322.322.47

[1]　AhujaRK,KumarA,JhaKC,etal.Exactandheuristicmethodsfortheweapontargetassignmentproblem[J].Opera-1146.tionsResearch,2007,55(6):1136-[2]　NiMingfang,YuZhanke,MaFeng,etal.ALagrangerelaxationmethodforsolvingweapon-targetassignmentproblem

[J].MathematicalProblemsinEngineering,2011,5(1):1-12.

[3]　CetinE,EsenST.Aweapon-targetassignmentapproachtomediaallocation[J].AppliedMathematicsandComputa-tion,2006,175(2):1266-1275.

[4]　WolseyLA,NemhauserGL.Integerandcombinatorialoptimization[M].NewYork:JohnWilley&Sons,1988:540-546.

[5]　NocedalJ,WrightSJ.Numericaloptimization[M].NewYork:Springer,1999:355-385.

(上接第22页)参考文献:

[1]　KamathS,LoizouP.Amulti-bandspectralsubtractionmethodforenhancingspeechcorruptedbycolorednoise[J].

IEEETransAcoust,SpeechSignalProcess,2002,8(4):164-168.

[2]　EphraimY,VanTreesHL.Asignalsubspaceapproachforspeechenhancement[J].IEEETransAcoust,SpeechSignal

Process,1995,3(4):251-266.

[3]　YiH,PhiliposCL.Ageneralizedsubspaceapproachforenhancingspeechcorruptedbycolorednoise[J].IEEETransA-coust,SpeechSignalProcess,2003,11(4):334-341.

[4]　SamTR.Onemicrophonesourceseparation[J].AdvancesinNeuralInformationProcessingSystems,2000,13(2):793-799.

[5]　DanielPW,RonJW.Model-basedmonauralsourceseparationusingavector-quantizedphase-vocoderrepresentation

[C]∥InternationalConferenceonAcoustics,SpeechandSignalProcessing.France:IEEEPress,2006:957-960.

[6]　陈善学,杜　锋.一种改进的等误差自组织特征映射矢量量化方法[J].重庆邮电大学学报:自然科学版,2011,23(2):

155-160.

[7]　LeeDD,SeungHS.Learningthepartsofobjectsbynon-negativematrixfactorization[J].Nature,1999,401(10):788-791.

[8]　FujisawaH,EguchiS.Robustparameterestimationwithasmallbiasagainstheavycontamination[J].MultivariateAnal-2081.ysis,2008,99(9):2053-[9]　KristianTA.Windnoisereductioninsinglechannelspeechsignals[D].Denmark:TechnicalUniversityofDenmark,

2008.

[10]MikkelN,SchmidtRK,Olsson.Single-channelspeechseparationusingsparsenon-negativematrixfactorization[C]∥

InternationalConferenceonSpokenLangnageProcessing(INTERSPEECH).Pittsburgh:UniversityofPittsburgh,2006:821-824.

[11]WangB,PlumbleyMD.Musicalaudiostreamseparationbynon-negativematrixfactorization[C]∥Proceedingsofthe

D