基于差分进化的改进狼群算法研究

Research of improved wolf pack algorithm based on differential evolution

针对传统狼群算法(WPA)存在易陷入局部最优解、计算资源耗费大、鲁棒性低等问题,提出一种基于差分进化的改进狼群算法(DWPA)。首先,通过引入探狼搜索因子、猛狼最大奔袭次数、自适应围攻步长、差分进化策略等对传统狼群算法进行了改进,在降低算法计算耗费的同时提高了算法的全局搜索能力;然后,运用马尔可夫链理论证明了DWPA的收敛性;最后,对13个测试函数进行寻优测试,并与WPA等四种算法进行对比分析。测试结果表明,DWPA具有良好的鲁棒性和全局搜索能力,在求解多峰、高维、不可分函数方面的寻优能力尤为突出。

Aiming at the problems of traditional wolf pack algorithm (WPA), such as easy to fall in to local optimal, large computational resource cost and low robustness,this paper proposed an improved wolf pack algorithm based on differential evolution (DWPA).First of all,it proposed search wolf search factor , maximum number of raid wolves,adaptive siege step size and differential evolution strategy to improve the traditional wolf pack algorithm,which could not only reduce the computational cost of the algorithm but also improved the global search ability.Then,it proved the convergence of DWPA applying the Markov process.Finally,it conducted optimization test on 13 functions and then compared it with WPA and other 4 algorithms. The test results show that DWPA has great robustness and global search ability, especially has an excellent optimizing ability in multi-peak, high-dimension, indivisible functions.