摘要
为改善蝙蝠算法求解高维函数优化问题的全局搜索能力,提高其搜索精度,将交叉熵方法和蝙蝠算法相结合,提出一种交叉熵蝙蝠算法。该算法将基于重要度抽样和Kullback-Leibler距离的交叉熵全局随机优化算法应用于蝙蝠算法中,采用自适应平滑技术提高算法的收敛速度,利用交叉熵方法的遍历性、自适应性和鲁棒性,有效抑制蝙蝠算法的早熟收敛现象。对经典测试函数和CEC2005测试函数的仿真结果表明,该算法具有全局搜索能力强、求解精度高和鲁棒性好等特性。
In order to enhance the global search ability of bat algorithm in solving high-dimensional function optimization problems,a Cross-entropy Bat Algorithm（ CEBA） is proposed by combining bat algorithm with CE method. The CE global stochastic optimization algorithm which is based on importance sampling and Kullback-Leibler divergence, is embedded into bat algorithm. By using adaptive smoothing technique, CEBA improves the rate of convergence. The improved algorithm fully absorbs the ergodicity,adaptability and robustness of CE,adaptively avoids the stagnancy of population,and enhances the global search ability. Simulated results conducted on classical benchmarks and 10 CEC2005 benchmarks show that the proposed algorithm possesses more powerful global search capacity,higher optimization precision and robustness.
出处
《计算机工程》
CAS
CSCD
2014年第10期168-174,180,共8页
Computer Engineering
基金
国家自然科学基金资助项目(11171221)
上海市一流学科(系统科学)基金资助项目(XTKX2012)
关键词
高维函数优化
蝙蝠算法
交叉熵
重要度抽样
自适应平滑
协同演化
high-dimensional function optimization
Bat Algorithm （BA）
Cross-entropy （ CE ）
importance sampling
adaptive smoothing
co-evolution
