Scholarlink

<p>This article presents an enhanced Landweber method with momentum acceleration for solving large indefinite least squares problem. Theoretically, we analyze the interval of values of the momentum parameter when this iterative method satisfies the convergence condition and derive the value of the optimal momentum parameter for the maximum convergence rate. Numerical experiments validate the theoretical findings and demonstrate that the Landweber method with momentum acceleration is superior to the USSOR method (Song, Int. J. Comput. Math. 97, 1781-1791 2020) and the splitting-based randomized iterative method (Zhang and Li, Appl Math Comput 446:127892, 2023) in terms of algorithmic running time.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

New iterative method with fast convergence of the momentum for solving indefinite least squares problem

  • 期刊论文

  • 发表日期：

    2025年6月18日

查看全文
  • Lingsheng Meng,
  • Kailiang Xin,
  • Yunying Huang

摘要

This article presents an enhanced Landweber method with momentum acceleration for solving large indefinite least squares problem. Theoretically, we analyze the interval of values of the momentum parameter when this iterative method satisfies the convergence condition and derive the value of the optimal momentum parameter for the maximum convergence rate. Numerical experiments validate the theoretical findings and demonstrate that the Landweber method with momentum acceleration is superior to the USSOR method (Song, Int. J. Comput. Math. 97, 1781-1791 2020) and the splitting-based randomized iterative method (Zhang and Li, Appl Math Comput 446:127892, 2023) in terms of algorithmic running time.