YYPO-SA:一种新的基于YYPO和SA的混合单目标随机优化算法

YYPO-SA:novel hybrid single-object optimization algorithm based onYin-Yang-pair optimization and simulated annealing

针对原阴阳对优化算法(YYPO)早熟易收敛的问题,在YYPO算法中的阴阳两点交换阶段加入模拟退火算法(SA)策略,提出了两种使用不同交换策略的新算法,即YYPO-SA1和YYPO-SA2,统称为YYPO-SA。YYPO-SA算法既保持了YYPO轻量级的特点,又综合了YYPO优秀的全局搜索能力和SA良好的局部搜索性能。算法采用2013年进化计算大会的单目标实参算法竞赛中使用的28个测试函数进行性能评估,将YYPO-SA和YYPO、自适应阴阳对算法(AYYPO)、改进的阴阳对算法(IYYPO),以及另三个性能优越的单目标优化算法,即灰狼优化算法、鲸鱼优化算法,正弦余弦算法进行性能比较。实验结果表明YYPO-SA能取得更为稳定的求优能力和更高的计算精度。最后通过一个工程优化任务来展示新算法的性能。

Aiming at solving the premature convergence of YYPO algorithm,this paper introduced SA strategy into the switch phase between P 1 and P 2 in YYPO,and proposed a novel hybrid single-object optimization algorithms YYPO-SA(including YYPO-SA1 and YYPO-SA2).YYPO-SA inherited the lightweight characteristics of YYPO,and combined the YYPO’s excellent global search ability with the powerful local search performance of SA.It used the proposed 28 test functions in the single objective practical parameter algorithm competition in 2013 Evolutionary Computing Conference as the benchmark.It empirically studied the performance of YYPO-SA2,YYPO,A-YYPO,IYYPO,and 3 state of the art algorithms——GWO,WOA,and SCA based on experimental results.The experimental results show that YYPO-SA can perform more stable and achieve more accurate solutions.At the end,this paper also applied YYPO-SA to solve the spring design task.The experimental results show that YYPO-SA can also solve the constrained optimization task well.