| 引用本文: | 洪露,龚成龙,王经卓,纪志成.噪声环境下精英克隆选择算法的收敛性分析[J].控制理论与应用,2013,30(11):1457~1461.[点击复制] |
| HONG Lu,GONG Cheng-long,WANG Jing-zhuo,JI Zhi-cheng.Convergence analysis of elitist clonal selection algorithm in noisy environment[J].Control Theory and Technology,2013,30(11):1457~1461.[点击复制] |
|
| 噪声环境下精英克隆选择算法的收敛性分析 |
| Convergence analysis of elitist clonal selection algorithm in noisy environment |
| 摘要点击 3175 全文点击 2319 投稿时间：2013-05-13 修订日期：2013-07-19 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2013.30471 |
| 2013,30(11):1457-1461 |
| 中文关键词 克隆选择算法 精英策略 加性噪声 鞅理论 转移概率 |
| 英文关键词 clonal selection algorithm elitist strategy additive noisy martingale theory transition probability |
| 基金项目 国家自然科学基金资助项目(61174013); 江苏高校优势学科建设工程资助项目. |
|
| 中文摘要 |
| 利用随机过程相关理论对加性噪声环境下精英策略克隆选择算法(ECSA)的全局收敛性进行了研究. 首先采用有序对的状态表示方法构造精英克隆选择算法在噪声环境中的Markov链; 然后将算法种群中最佳亲和度函数的进化过程转化为下鞅, 利用鞅理论证明了种群最佳亲和度函数的全局收敛性; 最后通过分析加性噪声环境下精英克隆选择算法的状态转移概率的特性, 证明了精英克隆选择算法在加性噪声环境下最终能以概率1 收敛到全局最优解. |
| 英文摘要 |
| The global convergence of elitist clonal selection algorithm (ECSA) in additive noisy environment is studied by using stochastic process. The Markov chain that models ECSA in noisy environment is constructed by a new state representation method named ordered pair. The evolutionary sequence best fitness functions is described as a submartingale, and the global convergence of the best fitness functions is derived based on martingale theory. The transition probability characteristics of the clonal selection algorithm in additive noisy environment are analyzed; by this property we conclude that ECSA is eventually globally convergent with probability 1 in additive noisy environment. |