多策略混合的天鹰优化器  

Aquila Optimizer with Multi-Strategy Integration

作　　者：刘香怡 梁宏涛 朱洁 LIU Xiangyi;LIANG Hongtao;ZHU Jie(School of Information Science and Technology,Qingdao University of Science and Techology,Qingdao 266011,China)

机构地区：[1]青岛科技大学信息科学技术学院,山东青岛266011

出　　处：《计算机测量与控制》2024年第8期295-303,共9页Computer Measurement &Control

基　　金：国家自然科学基金项目(61973180;62172249)。

摘　　要：为了解决天鹰优化器集中在全局搜索导致的局部寻优能力略差、依赖初始种群质量和易陷入局部最优的问题,提出一种多策略混合的天鹰优化器;该算法利用改进的Hooke-jeeves优化基本天鹰优化器的初始化种群质量;引入模拟退火概率对易陷入局部最优解进行改进;自适应权重提高前期全局搜索效率,延缓后期局部搜索速度,避免在正解附近徘徊;选取12个基准测试函数进行实验,并将MAO应用于风力发电预测模型优化;实验结果表明,对于单峰函数、多峰函数和固定维函数,MAO比AO等对比函数具有更快的收敛速度和更高的精度;在春夏秋冬数据集上进行仿真实验,对比其他模型1月和10月预测精度提高了15%,4月和8月的预测曲线更加平滑;证实了MAO对于提高风电预测的精度和速度的可行性和实用性。In order to solve the problems of poor local optimization ability,dependence on the quality of initial population,and easily falling into local optimum caused by the global search of aquila optimizer(AO),a multi-strategy integration AO is proposed.The algorithm utilizes the improved Hooke-jeeves alogrithm to optimize the initialized population quality of the basic aquila optimizer.The simulated annealing probability is introduced to improve the local optimal solution,The adaptive weights improve the efficiency of the global search in the early stage and slow down the local search in the late stage to avoid hovering around the positive solution.Through selecting 12 benchmark test functions for experiments,and the mixed aquila optimizer(MAO) is applied to optimize the wind power prediction model.Experimental results show that for single-peak,multi-peak and fixed-dimension functions,the MAO has faster convergence speed and higher accuracy than comparative functions such as the AO.Simulation experiments are implemented on spring,summer,fall and winter datasets,compared with other models,the prediction accuracy in January and October is improved by 15%,and the prediction curves in april and august are smoother.It is verified that the MAO improves the feasibility and practicability of wind power prediction accuracy and speed.

关 键 词：天鹰优化器 Hooke-Jeeves算法 模拟退火 自适应权重 风电预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TM614[自动化与计算机技术—控制科学与工程]