【作者】 夏银华；

【导师】 舒其望；

【作者基本信息】 中国科学技术大学 ， 计算数学， 2008， 博士

【摘要】 本文中,我们构造了四阶完全非线性Cahn-Hilliard(CH)型方程和Allen-Cahn/Cahn-Hilliard(AC/CH)方程组的局部间断有限元方法,并证明了一般非线性情形下,CH型方程和AC/CH方程组的LDG方法在任意维数、网格和精度下的能量稳定性。该方法具有高精度,非线性稳定以及h-p自适应的优点。关于一维和二维空间下的CH型方程和AC/CH方程组的数值实验说明了LDG方法的精度和能力,表明了LDG方法在求解这类完全非线性高阶方程时十分有效。我们构造了三角形网格下行人流动态反应行人平衡模型的DG算法。该模型由关于行人流密度的守恒律方程控制,并且流通量的方向通过Eikonal方程隐式依赖于行人流密度。在数值求解过程中,每一时间步内我们利用三角形网格的快速扫描算法求解Eikonal方程,并通过求解在两个有不同障碍物的站台内的行人流算例来说明该算法的能力,从而快速有效地解决了任意几何区域上行人流模型的求解问题。为了提高求解LDG空间离散具有高阶导数的偏微分方程后得到的常微分方程组的效率,我们研究了三种半隐时间离散方法。由于LDG空间离散后的半离散形式常微分方程组在时间上呈现多尺度,因此刚性部分要求显式时间离散的时间步长非常小。我们将着重讨论半隐谱延迟校正(SDC)时间离散方法,以及结合LDG空间离散算子后它的稳定性和精度。同时我们也将讨论另外两种时间离散方法,additive Runge-Kutta(ARK)方法和指数时间离散(ETD)方法。并比较三种方法结合LDG空间离散方法在求解含有高阶导数方程时的效率。我们看到三种半隐方法相比显式时间离散方法都是高效的。要指出的是,SDC方法可以轻易推广到任意阶精度,而ARK方法在我们的数值试验中所需的CPU时间最少。更多还原

【Abstract】 In this thesis we develop local discontinuous Galerkin(LDG)methods for the fourth-order fully nonlinear Cahn-Hilliard(CH)type equations and Allen-Cahn/Cahn-Hilliard(AC/CH)system.The energy stability of the LDG meth-ods is proved for any orders of accuracy on arbitrary triangulations in any space dimension for the general nonlinear case.The LDC discretization results in high order accuracy,nonlinear stable and suitable for hp-adaptation scheme.Numeri-cal examples for the Cahn-Hilliard equations and the Allen-Cahn/Cahn-Hilliard system in one and two dimensions are presented and the numerical results illus-trate the accuracy and capability of the methods.These results indicate that the LDC method is a good tool for solving such nonlinear equations in mathematical physics.We also develop a discontinuous Galerkin method on triangular meshes to solve the reactive dynamic user equilibrium model for pedestrian flows.The pedestrian density in this model is governed by the conservation law in which the flow flux is implicitly dependent on the density through the EikonaI equation.To solve the Eikonal equation efficiently at each time level,we use the fast sweeping method.Two numerical examples are then used to demonstrate the effectiveness of the algorithm.This algorithm efficiently solves the pedestrian flows problem on arbitrary geometry domain.We explore three efficient time discretization techniques for the LDG meth-ods to solve partial differential equations(PDEs)with higher order spatial deriv-atives.The main difficulty is the stiffness of the LDC spatial discretization op-erator,which would require a unreasonably small time step for an explicit local time stepping method.We focus our discussion on the semi-implicit spectral deferred correction(SDC)method,and study its stability and accuracy when coupled with the LDG spatial discretization.We also discuss two other time discretization techniques,namely the additive Runge-Kutta(ARK)method and the exponential time differencing(ETD)method,coupled with the LDG spatial discretization.A comparison is made among these three time discretization tech-niques,to conclude that all three methods are efficient when coupled with the LDG spatial discretization for solving PDEs containing higher order spatial deriv-atives.In particular,the SDC method has the advantage of easy implementation for arbitrary order of accuracy,and the ARK method has the smallest CPU cost in our implementation.更多还原