Abstract: Aiming at the problems of poor convergence, low searching precision and ease of premature convergence when solving the complex optimization problems, combining with adaptive strategy, an improved SOS algorithm with different difference perturbation terms and elite opposition-base learning strategy is proposed. Experiments are conducted on the 14 benchmark functions and the results show that the improved SOS algorithm has obviously better performance in convergence speed, solution precision and global optimization than SOS algorithm and other three algorithms.

Key words: Symbiotic Organisms Search（SOS）, difference perturbation, adaptive adjustment, elite opposition-based learning

摘要： 针对共生生物搜索算法在求解高维复杂问题时存在过早收敛，求解精度不高及后期搜索迟滞等问题，结合自适应思想，利用不同差分扰动项和精英反向学习策略对算法进行改进，得到一种改进的共生生物搜索算法。对14个标准测试函数的仿真实验结果进行分析，相比于原算法和其他三种目前流行的算法，改进算法在收敛速度和求解精度方面均具有明显的优势，寻优能力更强。

关键词: 共生生物搜索算法, 差分扰动, 自适应, 精英反向学习