计算机工程与应用
計算機工程與應用
계산궤공정여응용
COMPUTER ENGINEERING AND APPLICATIONS
2009年
31期
49-52
,共4页
微粒群算法%马氏过程%函数优化
微粒群算法%馬氏過程%函數優化
미립군산법%마씨과정%함수우화
particle swarm optimization%markov process%function optimization
受遗传算法马氏模型理论分析的启发,提出了一种便于用马氏过程理论分析的微粒群算法.该算法中的个体仅记忆群体在进化过程中有限步内的信息.忘掉以前的信息,以建立算法的马氏过程数学模型.通过函数优化的数值模拟验证了新算法具备优良的寻优能力,同时论证了新算法是齐次马氏过程.
受遺傳算法馬氏模型理論分析的啟髮,提齣瞭一種便于用馬氏過程理論分析的微粒群算法.該算法中的箇體僅記憶群體在進化過程中有限步內的信息.忘掉以前的信息,以建立算法的馬氏過程數學模型.通過函數優化的數值模擬驗證瞭新算法具備優良的尋優能力,同時論證瞭新算法是齊次馬氏過程.
수유전산법마씨모형이론분석적계발,제출료일충편우용마씨과정이론분석적미립군산법.해산법중적개체부기억군체재진화과정중유한보내적신식.망도이전적신식,이건립산법적마씨과정수학모형.통과함수우화적수치모의험증료신산법구비우량적심우능력,동시론증료신산법시제차마씨과정.
Inspired by the theoretic analysis of genetic algorithm based on markov process,a new form of particle swarm optimization algorithm is advanced,which is convenient for analysis by the theory of markov process.The particle of new algorithm only memorizes the information of swarm in finite steps,and forgets the old information.Then the markov process model is established/The simulations of functions optimization show that the new algorithm has good ability to find the global solution,and the homogeneous markov process is got from the new algorithm.
