摘要
针对鸽群优化算法容易陷入局部最优、求解精度低和局部搜索能力差的问题,提出了一种融合Tent映射和模拟退火的改进鸽群优化算法。首先,采用Tent映射初始化种群,使初始种群分布更为均匀。然后,在鸽群优化算法的地标算子运行后再加入模拟退火算法,利用模拟退火算法以一定的概率跳出局部最优解以及具有的渐进收敛性,提高了全局优化的能力。基于16个基准测试函数对算法性能进行了测试,实验结果表明,Tent-PIO-SA算法相比PIO、Tent-PIO算法,在收敛精度上平均提高了10个数量级,特别对于极难优化的Rosenbrock函数,Tent-PIO-SA算法相比最近的经典算法LECUSSA、SCASL、CML-WOA、APN-WOA在收敛精度上平均高出了6个数量级,比TLPSO、SCA-DE算法高出了7个数量级,证明了所提出的Tent-PIO-SA算法具有较强的寻优能力。
Aiming at the problems that pigeon swarm optimization algorithm is easy to fall into local optimum,low solution accuracy and the poor local searching ability,improved pigeon swarm optimization algorithm combining Tent mapping and simulated annealing was proposed.Firstly,Tent mapping was used to initialize the population to make the initial population distribution more uniform.Then,the simulated annealing algorithm was added after running the landmark operator of the pigeon swarm optimization algorithm.The simulated annealing algorithm can jump out of the local optimal solution with a certain probability and has the asymptotic convergence to improve the ability of global optimization.The experimental results on the performance of the algorithm based on 16 benchmark functions show that the convergence accuracy of Tent-PIO-SA algorithm is improved by an average of 10 orders of magnitude compared with PIO and Tent-PIO algorithms.Especially for Rosenbrock function,which is extremely difficult to optimize,the convergence accuracy of Tent-PIO-SA algorithm achieves on average 6 orders of magnitude higher than the recent classical algorithms LECUSSA,SCASL,CML-WOA and APN-WOA,and 7 orders of magnitude higher than TLPSO,SCA-DE algorithm.
作者
张安玲
ZHANG Anling(Department of Mathematics,Changzhi University,Changzhi 046011,China)
出处
《中北大学学报(自然科学版)》
2025年第1期53-63,75,共12页
Journal of North University of China(Natural Science Edition)
关键词
鸽群优化算法
模拟退火算法
TENT映射
pigeon swarm optimization algorithm
simulated annealing algorithm
Tent mapping
