欢迎来到天天文库
浏览记录
ID:18653319
大小:676.51 KB
页数:9页
时间:2018-09-20
《全信息变权缓冲算子的构造及应用》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、第47卷第6期浙江大学学报(工学版)Vol.47No.62013年6月JournalofZhejiangUniversity(EngineeringScience)Jun.2013DOI:10.3785/j.issn.1008-973X.2013.06.027全信息变权缓冲算子的构造及应用王正新(浙江财经大学经济与国际贸易学院,浙江杭州310018)摘要:针对现有变权缓冲算子简单的结构导致其不能充分利用系统行为信息的问题,基于新信息优先原理,构造一类带有权重调节因子的全信息变权弱化缓冲算子和强化缓冲算子.考虑到权重调节因子与预测误差之间的非线性关系,以平均预测误差最小化为优化目标,应用遗传算法
2、在(0,1)区间内搜索权重调节因子的最优值.以我国工业企业总产值和浙江省外商直接投资预测问题为例,分别验证了全信息变权弱化缓冲算子和强化缓冲算子处理小样本扰动数据建模问题的有效性.结果表明:全信息变权缓冲算子对序列的作用过程体现了新信息优先原理,通过优化权重调节因子能够有效地控制算子作用强度,且对冲击扰动系统的预测能力优于传统的高阶缓冲算子.关键词:灰色系统;新信息优先;变权缓冲算子;作用强度;经济预测中图分类号:N941.5文献标志码:A文章编号:1008-973X(2013)06-1120-09Constructionandapplicationofvariable-weightbuffe
3、roperatorswithperfectinformationWANGZheng-xin(SchoolofEconomicsandInternationalTrade,ZhejiangUniversityofFinanceandEconomics,Hangzhou310018,China)Abstract:Todealwiththeproblemoftheinsufficientutilizationofthebehaviorsystems’informationduetothesimplestructuresofexistingvariable-weightbufferoperators,
4、aclassofvariableweightsweakeningandstrengtheningbufferoperatorswithweights’regulatingfactorsandperfectinformationwereconstructedbasedonthenewinformationpriorityprinciple.Consideringthenonlinearrelationshipsbetweentheweights’regulatingfactorsandpredictionerrors,thegeneticalgorithmwasemployedtosearcht
5、heoptimalregulatingfactorininterval(0,1),undertheoptimizationobjectiveofminimizingtheaveragepredictionerror.ThepreditionproblemsofthegrossoutputofindustrialenterprisesinChinaandtheforeigndirectinvestmentinZhejiangProvinceofChinawerestudiedtoverifytheeffectivenessofthevariable-weightbufferoperatorswi
6、thperfectinformationfordealingwiththemodelingproblemofsmallsampledisturbeddata.Theresultsshowthattheactionprocessofthenewbufferoperatorsaffectingoriginaldataembodiesthenewinformationpriorityprinciple.Theoptimizationoftheweightregulatingfactorscaneffectivelycontroltheinteractionstrength.Theforecastin
7、gabilityofthenewbufferoperatorsfordisturbedsystemsissuperiortothetraditionalhighorderbufferoperators.Keywords:greysystem;newinformationpriority;variable-weightbufferoperator;interactionstrength;econom
此文档下载收益归作者所有