摘要
将变异算子与Bayesian优化算法相结合,提出了一种基于变异的Bayesian优化算法。在算法中设计了一个种群多样性函数,通过此函数引入变异算子,目的是利用变异算子的邻域搜索能力,保持种群多样性,将贝叶斯概率模型提取的全局信息与变异算子的局部信息联系起来,避免陷入局部最优。仿真研究表明基于变异的Bayesian优化算法的寻优能力比Bayesian优化算法更强。
A new Bayesian optimization algorithm is presented by incorporating mutation operator into Bayesian optimization algorithm. A diversity function of population is proposed and the mutation operator is incorporated in BOA through this function. The original objective is to maintain the diversity of population using the neighborhood search of mutation operator. It is expected that the proposed algorithm can get genuine global information by combining the global information in current population extracted by Bayesian probability model and local information explond by mut,~tion operator. Experimental results show that the proposed algorithm outperforms BOA.
出处
《计算机工程》
CAS
CSCD
北大核心
2007年第16期153-155,158,共4页
Computer Engineering
基金
国家自然科学基金资助项目(60374063)
