a cheyshev-gauss-lobatto spectral collocation method for nonlinear volterra integral equations with vanishing variable delays.pdfVIP

Shanghai Normal University Master Dissertation 摘 要 谱方法是求解微分方程的一种重要数值方法，已被广泛应用于科学和工程问题的数值 模拟中，其主要优点是计算的高精度。另一方面,Volterra时滞积分方程具有记忆性质，在 物理、生物、激光以及人口增长等模型中得到广泛应用，相关的数值研宄正日益受到重 视，并已成为该领域的一个新热点。 现有的针对Volterra时滞方程谱方法的研宄主要基于单步格式，并不适合长时间的计 算。此外，所研宄的问题主要是线性的，而实际问题大多是非线性的。因此有必要研宄非 线性 Volterra型时滞积分方程的多步谱方法。 本文主要研宄非线性消失时滞Volterra积分方程的多步Chebyshev-Gauss-Lobatto谱配 置方法。我们建立了相关问题的多步谱配置格式，并给出了相应的算法。数值结果表明， 所提方法具有高精度，长时间计算稳定，且对于高振荡问题、局部大梯度问题以及非光滑 解问题等十分有效。 本文由以下三个部分组成： 第一章，我们简单地回顾了时滞Volterra积分方程数值方法的研宄进展。 第二章，我们提出了非线性消失时滞 Volterra积分方程的多步 Chebyshev-Gauss- Lobatto谱配置方法，并给出了相应的算法。 第三章，通过数值算例，验证本文所提出的数值格式的有效性。 关键词：多步Chebyshev-Gauss-Lobatto配置法，非线性消失时滞Volterra积分方程，数值 试验。 I 万方数据 Shanghai Normal University Master Dissertation Abstract As an important numerical method, spectral method often provides exceedingly accurate nu merical results with relatively less degree of freedoms, and has been widely used for scientific computations. On the other hand, Volterra integral equations with delays (DVIEs) are the models of evolutionary problems with memory arising in many applications, such as physical and biolog ical phenomena, lasers and population growth, etc.. In recent years, some numerical approaches for such problems have become increasingly popular. Among the existing works, numerical methods based on the single-step spectral schemes have been frequently used for various linear Volterra-type equations. However, this kind of algorithms is not suitable for long time simulations. Therefore, it is also very important for us to study the multistep spectral methods of the nonlinear Volterra-type equations. The main purpose of this dissertation is to investigate the multistep Chebyshev-Gauss-Lobatto