西安电子科技大学学报(自然科学版)
西安電子科技大學學報(自然科學版)
서안전자과기대학학보(자연과학판)
JOURNAL OF XIDIAN UNIVERSITY(NATURAL SCIENCE)
2014年
4期
186-192
,共7页
网格%代理%多重链接%节点权重
網格%代理%多重鏈接%節點權重
망격%대리%다중련접%절점권중
grid%agent%multilink%nodal weight
将制造网格与Web技术相结合建立了一个新的制造网格控制模型,可有效地刻画制造网格的有向控制过程,同时具有易管理和易控制的特征。在构建模型过程中给出了状态汇聚代理的概念,并分析了状态汇聚代理在有向制造网格进行协同控制管理的过程中提高管理效果和可控性的工作机理。基于该工作机理提出一种基于多重链接的有向制造网格节点权重计算方法,可根据节点权重计算结果排序来识别有向制造网格中的状态汇聚代理。采用T检验、Kendall-tau和斯皮尔曼等级相关系数这3种统计学检验标准对不同的状态汇聚代理识别算法进行对比实验,同时还对算法的准确率和召回率以及计算时间进行了实验验证。结果表明,从状态汇聚代理识别能力、准确率和召回率以及计算时间等综合指标来看,新方法更具优势。
將製造網格與Web技術相結閤建立瞭一箇新的製造網格控製模型,可有效地刻畫製造網格的有嚮控製過程,同時具有易管理和易控製的特徵。在構建模型過程中給齣瞭狀態彙聚代理的概唸,併分析瞭狀態彙聚代理在有嚮製造網格進行協同控製管理的過程中提高管理效果和可控性的工作機理。基于該工作機理提齣一種基于多重鏈接的有嚮製造網格節點權重計算方法,可根據節點權重計算結果排序來識彆有嚮製造網格中的狀態彙聚代理。採用T檢驗、Kendall-tau和斯皮爾曼等級相關繫數這3種統計學檢驗標準對不同的狀態彙聚代理識彆算法進行對比實驗,同時還對算法的準確率和召迴率以及計算時間進行瞭實驗驗證。結果錶明,從狀態彙聚代理識彆能力、準確率和召迴率以及計算時間等綜閤指標來看,新方法更具優勢。
장제조망격여Web기술상결합건립료일개신적제조망격공제모형,가유효지각화제조망격적유향공제과정,동시구유역관리화역공제적특정。재구건모형과정중급출료상태회취대리적개념,병분석료상태회취대리재유향제조망격진행협동공제관리적과정중제고관리효과화가공성적공작궤리。기우해공작궤리제출일충기우다중련접적유향제조망격절점권중계산방법,가근거절점권중계산결과배서래식별유향제조망격중적상태회취대리。채용T검험、Kendall-tau화사피이만등급상관계수저3충통계학검험표준대불동적상태회취대리식별산법진행대비실험,동시환대산법적준학솔화소회솔이급계산시간진행료실험험증。결과표명,종상태회취대리식별능력、준학솔화소회솔이급계산시간등종합지표래간,신방법경구우세。
A novel control model of the manufacturing grid based on Web technology is established.It describes the directed control process of the manufacturing grid and has the features of manageability and controllability.A notion of status collection agent is proposed in the process of model establishment.And the mechanism in which status collection agents improve the management efficiency and controllability in manufacturing grid cooperative control of the management process is introduced. According to the characteristics of the directed manufacturing grid control model,a method for computing the weight of the directed manufacturing grid node based on the multilink is proposed.Status collection agents are identified in the light of former sort results.Contrastive experiments on different recognized algorithms of status collection agents are conducted by adopting T-Test,Kendall-tau and Spareman Rank statistic tests,and in the meantime,experiments have been performed on the accurate rate and recall rate of the algorithm and calculating time. The results show that this method is better from comprehensive specifications of recognition capability on status collection agents,accurate rate,recall rate and computing time.