摘要
针对灰狼优化算法不足,提出融合梯度模拟权重和自适应差分的灰狼优化算法.首先,设计多收敛因子以增加全局勘探能力,且维持局部开发能力;其次,利用分组方式将概率自适应差分策略与梯度模拟的动态权重策略进行结合,前者具有增加种群多样性和全局优化的能力,后者有效提高收敛速度和精度;最后,设计反向学习的高斯-柯西变异以增加算法的跳出局部最优能力.通过与5个先进的智能优化算法、3个灰狼变体算法比较以及消融实验方法,验证该算法的有效性和先进性.16个经典测试函数和优化Richards模型参数问题上的实验结果表明,该算法在收敛速度和收敛精度上具有显著优势,且在参数优化方面也有效.
Aiming at the limitation of the gray wolf optimization algorithm,a gray wolf optimization algorithm integrating gradient simulation weights and adaptive differencing is proposed.Firstly,multiple convergence factors are designed to enhance the global exploration ability and maintain the local exploitation ability.Then,the probabilistic adaptive difference strategy is combined with the dynamic weighting strategy of gradient simulation by grouping.The former can increase the diversity of populations and global optimization,and the latter effectively improves the convergence speed and accuracy.Finally,the Gaussian-Cauchy variant of inverse learning is designed to improve the ability of the algorithm to jump out of the local optimum.The effectiveness and sophistication of GDGWO are verified by using ablation experimental method and comparing with five advanced intelligent optimization algorithms and three gray wolf variant algorithms.The experimental results on 16 classical test functions and optimizing Richards model parameter problems show that GDGWO has significant advantages in convergence speed and convergence accuracy,and it is also effective in parameter optimization problems.
作者
徐世杰
高岳林
XU Shijie;GAO Yuelin(School of Mathematics and Information Science,North Minzu University,Yinchuan Ningxia 750021;Ningxia Key Laboratory of Intelligent Information and Big Data Processing,Yinchuan Ningxia 750021)
出处
《宁夏师范大学学报》
2026年第1期55-68,共14页
Journal of Ningxia Normal University
基金
宁夏自然科学基金重点项目(2022AAC02043)
北方民族大学研究生创新项目(YCX24290).
关键词
灰狼优化算法
双收敛因子
自适应差分
梯度模拟
Richards模型
Gray wolf optimization algorithm
Double convergence factor
Adaptive differencing
Gradient simulation
Richards model
