中国机械工程
中國機械工程
중국궤계공정
CHINA MECHANICAl ENGINEERING
2014年
21期
2930-2935,2936
,共7页
陈东宁%姚成玉%王斌%张瑞星
陳東寧%姚成玉%王斌%張瑞星
진동저%요성옥%왕빈%장서성
可靠性优化%贝叶斯网络%引斥力%LRPSO 算法
可靠性優化%貝葉斯網絡%引斥力%LRPSO 算法
가고성우화%패협사망락%인척력%LRPSO 산법
reliability optimization%Bayesian network%attraction and repulsion%later-stage re-pulsion-enhanced hybrid attraction and repulsion particle swarm optimization(LRPSO)algorithm
将利用贝叶斯网络构造的系统故障概率函数作为可靠性指标，考虑费用、质量、体积构造了资源约束函数。针对微粒群算法引斥力规则的不足，提出了搜索后期斥力增强型混合引斥力微粒群算法(LRPSO 算法)：在搜索前期，使微粒在其他微粒的引斥力作用下进行最优搜索，以保持种群多样性；在搜索后期，减小引力、增强斥力，利用斥力项避免微粒陷入较差位置，以提高局部搜索能力。算法测试和可靠性优化实例验证了LRPSO 算法的有效性。
將利用貝葉斯網絡構造的繫統故障概率函數作為可靠性指標，攷慮費用、質量、體積構造瞭資源約束函數。針對微粒群算法引斥力規則的不足，提齣瞭搜索後期斥力增彊型混閤引斥力微粒群算法(LRPSO 算法)：在搜索前期，使微粒在其他微粒的引斥力作用下進行最優搜索，以保持種群多樣性；在搜索後期，減小引力、增彊斥力，利用斥力項避免微粒陷入較差位置，以提高跼部搜索能力。算法測試和可靠性優化實例驗證瞭LRPSO 算法的有效性。
장이용패협사망락구조적계통고장개솔함수작위가고성지표，고필비용、질량、체적구조료자원약속함수。침대미립군산법인척력규칙적불족，제출료수색후기척력증강형혼합인척력미립군산법(LRPSO 산법)：재수색전기，사미립재기타미립적인척력작용하진행최우수색，이보지충군다양성；재수색후기，감소인력、증강척력，이용척력항피면미립함입교차위치，이제고국부수색능력。산법측시화가고성우화실례험증료LRPSO 산법적유효성。
Fault probability function constructed by Bayesian network was regarded as reliability index,and the resource constraint functions were established by considering the functions of cost, weight and volume.To overcome the shortages of attraction and repulsion rule of particle swarm opti-mization algorithm,a LRPSO algorithm was proposed.At the earlier-stage,each particle searched the optimum under the attraction and repulsion produced by all particles,to maintain the population diversity.At the later-stage,the effect of attraction was reduced and the effect of repulsion was en-hanced using the repulsion term to avoid particles being trapped in worse searching position and im-proving local searching ability.The effectiveness of LRPSO algorithm was verified by algorithm tests and reliability optimization examples.
