资源描述：
《一种基于证据推理的装备保障资源评估方法》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。
1、第33卷第6期控制与决策Vol.33No.62018年6月ControlandDecisionJun.2018文章编号:1001-0920(2018)06-1048-07DOI:10.13195/j.kzyjc.2017.0288一种基于证据推理的装备保障资源评估方法周志杰1,刘涛源1,李方志2,赵福均1,胡冠宇3,于传强4(1.火箭军工程大学控制工程系，西安710025；2.96602部队北京102488；3.海南师范大学信息科学技术学院，海口570100；4.火箭军工程大学发射工程系，西安710025)摘要:为解决目前多属性决策分析在处
2、理不确定性问题上存在的不足,提出一种基于证据推理的装备保障资源评估方法,该方法通过融合多种不确定信息得到装备保障资源水平.首先,构建装备保障资源评估指标体系,根据专家知识获取指标权重;然后,基于规则的输入信息转化法,将指标数据转换到统一的置信框架下,并建立基于证据推理的评估模型.针对模型参数难以精确主观确定的困难,建立非线性优化模型,基于投影协方差矩阵自适应进化策略算法求解得到最优参数.对某单位装备保障资源实例进行评估,并通过对比分析验证所提出方法的有效性.关键词:装备保障资源；证据推理；评估；指标体系；参数优化中图分类号:TP181文献标
3、志码:AAnevaluationmethodofequipmentsupportresourcesbasedonevidentialreasoningZHOUZhi-jie1,LIUTao-yuan1,LIFang-zhi2,ZHAOFu-jun1,HUGuan-yu3,YUChuan-qiang4(1.DepartmentofControlEngineering，RocketForceUniversityofEngineering，Xi’an710025，China；2.TroopsofNo.96602，Beijing102488，Chi
4、na；3.CollegeofInformationScienceandTechnology，HainanNormalUniversity，Haikou570100，China；4.DepartmentofLaunchEngineering，RocketForceUniversityofEngineering，Xi’an710025，China)Abstract:Inordertosolvetheshortcomingsofthecurrentmulti-attributedecision-makinganalysisindealingwit
5、htheuncertainty,thispaperproposesanevaluationmethodofequipmentsupportresourcesbasedonevidentialreasoning.Toobtaintheevaluationlevel,varioustypesofinformationunderuncertaintyisintegrated.Firstly,theevaluationindexsystemofequipmentsupportresourcesisestablishedandtheweightsar
6、eobtainedfromexpertknowledge.Then,thedataofindexistransformedintoanunifiedbeliefframeworkusingtherule-basedtransformationtechniques,andtheevaluationmodelisdevelopedbasedonevidentialreasoning.Aimingatthedifficultyofsubjectivedeterminationofmodelparameters,anonlinearoptimizatio
7、nmodelisproposedandtheoptimalparametersareobtainedbasedontheprojectioncovariancematrixadaptationevolutionstrategy.Apracticalevaluationandcomparativeanalysisexampleofaunit’sequipmentsupportresourcesarestudiedtovalidatetheeffectivenessoftheproposedevaluationmethod.Keywords:eq
8、uipmentsupportresources；evidentialreasoning；evaluation；indexsystem；parametersoptimization
