摘要
提出了一种改进的Barzilai-Borwein共轭梯度法.选用一种新的初始步长策略,提出了一种广义Wolfe非单调线搜索,在函数f满足假设条件的前提下,建立了其全局收敛性和R-线性收敛性.将改进后的Barzilai-Borwein共轭梯度法应用于图划分问题,用改进算法求解了图划分问题中的无约束目标函数,并在相同的计算机环境中进行仿真实验.实验结果表明,改进算法能得到比原Barzilai-Borwein共轭梯度法更高质量的解.
An improved Barzilai-Borwein conjugate gradient method has been proposed.Firstly,a new initial step size strategy is selected,and a generalized Wolfe non-monotonic line search is proposed,and its global convergence and R-linear convergence are established under the premise that the function satisfies the assumptions.Finally,the improved Barzilai-Borwein conjugate gradient method is applied to the graph partitioning problem,and the improved algorithm is used to solve the unconstrained objective function in the graph partition problem,Simulation experiments are conducted in the same computer environment,and the experimental results show that the improved algorithm can obtain higher quality solutions than the original Barzilai-Borwein conjugate gradient method.
作者
吕佳敏
刘红卫
李瑶
游海龙
LYU Jiamin;LIU Hongwei;LI YaO;YOU Hailong(School of Mathematics and Statistics,Xidian University,Xi'an 710126,China;School of Microelectronics,Xidian University,Xi'an 710071,China)
出处
《东北师大学报(自然科学版)》
北大核心
2025年第2期45-55,共11页
Journal of Northeast Normal University(Natural Science Edition)
基金
国家自然科学基金资助项目(12261019).
