内容简介
This edition contains four new sections on the following topics: the BDDC domain decomposition preconditioner (Section 7.8), a convergent adaptive algorithm (Section 9.5), interior penalty methods (Section 10.5) and Poincare-Friedrichs inequalities for piecewise Wp1 functions (Section 10.6).We have made improvements throughout the text, many of which were suggested by colleagues, to whom we are grateful. New exercises have been added and the list of references has also been expanded and updated.
内页插图
目录
series preface
preface to the third edition
preface to the second edition
preface to the first edition
0 basic concepts
0.1 weak formulation of boundary value problems
0.2 ritz-galerkin approximation
0.3 error estimates
0.4 piecewise polynomial spaces - the finite element method
0.5 relationship to difference methods
0.6 computer implementation of finite element methods
0.7 local estimates
0.8 adaptive approximation
0.9 weighted norm estimates
0.x exercises
1 sobolev spaces
1.1 review of lebesgue integration theory
1.2 generalized (weak) derivatives
1.3 sobolev norms and associated spaces
1.4 inclusion relations and sobolev's inequality
1.5 review of chapter 0
1.6 trace theorems
1.7 negative norms and duality
1.x exercises
2 variational formulation of elliptic boundary value problems
2.1 inner-product spaces
2.2 hilbert spaces
2.3 projections onto subspaces
2.4 riesz representation theorem
2.5 formulation of symmetric variational problems
2.6 formulation of nonsymmetric variational problems
2.7 the lax-milgram theorem
2.8 estimates for general finite element approximation
2.9 higher-dimensional examples
2.x exercises
3 the construction of a finite element space
3.1 the finite element
3.2 triangular finite elements
the lagrange element
the hermite element
the argyris element
3.3 the interpolant
3.4 equivalence of elements
3.5 rectangular elements
tensor product elements
the serendipity element
3.6 higher-dimensional elements
3.7 exotic elements
3.x exercises
4 polynomial approximation theory in sobolev spaces
4.1 averaged taylor polynomials
4.2 error representation
4.3 bounds for riesz potentials
4.4 bounds for the interpolation error
4.5 inverse estimates
4.6 tensor. product polynomial approximation
4.7 isoparametric polynomial approximation
4.8 interpolation of non-smooth functions
4.9 a discrete sobolev inequality
4.x exercises
5 n-dimensional variational problems
5.1 variational formulation of poisson's equation
5.2 variational formulation of the pure neumann problem
5.3 coercivity of the variational problem
5.4 variational approximation of poisson's equation
5.5 elliptic regularity estimates
5.6 general second-order elliptic operators
5.7 variational approximation of general elliptic problems
5.8 negative-norm estimates
5.9 the plate-bending biharmonic problem
5.x exercises
6 finite element multigrid methods
6.1 a model problem
6.2 mesh-dependent norms
6.3 the multigrid algorithm
6.4 approximation property
6.5 w-cycle convergence for the kth level iteration
6.6 ]/-cycle convergence for the kth level iteration
6.7 full multigrid convergence analysis and work estimates
6.x exercises
7 additive schwarz preconditioners
7.1 abstract additive schwarz framework
7.2 the hierarchical basis preconditioner
7.3 the bpx preconditioner
7.4 the two-level additive schwarz preconditioner
7.5 nonoverlapping domain decomposition methods
7.6 the bps preconditioner
7.7 the neumann-neumann preconditioner
7.8 the bddc preconditioner
7.x exercises
8 max-norm estimates
8.1 main theorem
8.2 reduction to weighted estimates
8.3 proof of lemma 8.2.6
8.4 proofs of lemmas 8.3.7 and 8.3.11
8.5 lp estimates (regular coefficients)
8.6 lp estimates (irregular coefficients)
8.7 a nonlinear example
8.x exercises
9 adaptive meshes
9.1 a priori estimates
9.2 error estimators
9.3 local error estimates
9.4 estimators for linear forms and other norms
9.5 a convergent adaptive algorithm
9.6 conditioning of finite element equations
9.7 bounds on the condition number
9.8 applications to the conjugate-gradient method
9.x exercises
10 variational crimes
10.1 departure from the framework
10.2 finite elements with interpolated boundary conditions
10.3 nonconforming finite elements
10.4 isoparametric finite elements
10.5 discontinuous finite elements
10.6 poincare-friedrichs inequalitites for piecewise w1p functions
10.x exercises
11 applications to planar elasticity
11.1 the boundary value problems
11.2 weak formulation and korn's inequality
11.3 finite element approximation and locking
11.4 a robust method for the pure displacement problem
11.x exercises
12 mixed methods
12.1 examples of mixed variational formulations
12.2 abstract mixed formulation
12.3 discrete mixed formulation
12.4 convergence results for velocity approximation
12.5 the discrete inf-sup condition
12.6 verification of the inf-sup condition
12.x exercises
13 iterative techniques for mixed methods
13.1 iterated penalty method
13.2 stopping criteria
13.3 augmented lagrangian method
13.4 application to the navier-stokes equations
13.5 computational examples
13.x exercises
14 applications of operator-interpolation theory
14.1 the real method of interpolation
14.2 real interpolation of sobolev spaces
14.3 finite element convergence estimates
14.4 the simultaneous approximation theorem
14.5 precise characterizations of regularity
14.x exercises
references
index
精彩书摘
We will take this opportunity to philosophize about some power-ful characteristics of the finite element formalism for generating discreteschemes for approximating the solutions to differential equations. Being based on the variational formulation of boundary value problems, it is quite systematic, handling different boundary conditions with ease; one simply re-places infinite dimensional spaces with finite dimensional subspaces. What results, as in (0.5.3), is the same as a finite difference equation, in keeping with the dictum that different numerical methods are usually more similarthan they are distinct. However, we were able to derive very quickly the convergence properties of the finite element method. Finally, the notation for the discrete scheme is quite compact in the finite element for mulation.This could be utilized to make coding the algorithm much more efficient if only the appropriate computer language and compiler were available. Thislatter characteristic of the finite element method is one that has not yet been exploited extensively, but an initial attempt has been made in the sys-tem fec (Bagheri, Scott & Zhang 1992). (One could also argue that finiteele ment practitioners have already taken advantage of this by developingtheir own "languages" through extensive software libraries of their own, but this applies equally well to the finite-difference practitioners.)
......
前言/序言
Mathematics is playing an ever more important role in the physical and biological sciences, provoking a blurring of boundaries between scientific disciplines and a resurgence of interest in the modern as well as the clas-sical techniques of applied mathematics. This renewal of interest, both inresearch and teaching, has led to the establishment of the series Texts in Applied Mathematics (TAM).
The development of new courses is a natural consequence of a high level of excitement on the research frontier as newer techniques, such asnumerical and symbolic computer systems, dynamical systems, and chaos,mix with and reinforce the traditional methods of applied mathematics.Thus, the purpose of this textbook series is to meet the current and future needs of these advances and to encourage the teaching of new courses.
TAM will publish textbooks suitable for use in advanced undergraduateand beginning graduate courses, and will complement the Applied Mathe-matical Sciences (AMS) series, which will focus on advanced textbooks andresearch-level monographs.
《现代数值计算方法与算法》 内容简介 本书系统地探讨了现代数值计算领域的核心理论、关键算法及其在工程与科学计算中的应用。全书结构严谨,内容涵盖了从基础的线性代数运算到复杂偏微分方程数值解法的全貌,力求在理论深度与实际应用之间取得最佳平衡。 本书的目标读者是高年级本科生、研究生以及从事数值模拟、数据分析和计算科学的专业研究人员。它不仅提供了扎实的理论基础,更侧重于算法的内在机制、稳定性和收敛性分析,以及如何在实际计算环境中高效地实现这些算法。 第一部分:数值分析基础与误差理论 本部分奠定整个数值计算的理论基石。首先,对实数运算、浮点数表示、机器精度和截断误差、舍入误差进行了详尽的阐述。重点讨论了误差的传播和累积效应,特别是如何通过合理的算法设计来控制误差的增长。随后,深入分析了函数逼近的数学工具,包括多项式插值(如牛顿插值、拉格朗日插值),并详细讨论了插值余项的性质,引导读者理解“完美”插值在实际中的局限性。 第二部分:线性方程组的数值求解 线性系统 $Ax=b$ 的求解是计算科学的基石。本章首先回顾了矩阵代数的基础知识,随后聚焦于两大类求解方法:直接法和迭代法。 在直接法部分,详细分析了高斯消元法及其对计算复杂度的影响。重点讨论了矩阵的分解技术,特别是LU分解、Cholesky分解(针对对称正定矩阵)和QR分解。每种分解方法都配有详细的算法步骤、误差分析和对特定矩阵结构的适用性讨论。 迭代法部分,本书着重探讨了求解大型稀疏线性系统的有效策略。讲解了雅可比迭代和高斯-赛德尔迭代的收敛性条件(如对角占优矩阵),并深入剖析了更高级的迭代方法,如共轭梯度法(CG)、最小残量法(MINRES)和双共轭梯度法(BiCGStab)。针对非对称系统,对预处理技术(如代数多重网格预处理、代数重构预处理)的重要性进行了详尽的论述,强调预处理器对加速收敛速度的关键作用。 第三部分:非线性方程与优化问题 本部分转向非线性的世界。对于单变量非线性方程 $f(x)=0$,本书详细比较了割线法、牛顿法、信赖域法等方法的优缺点,特别是牛顿法在局部二次收敛性上的优势和计算成本。 在多变量非线性方程组 $mathbf{F}(mathbf{x})=mathbf{0}$ 的求解中,重点介绍了牛顿法及其准牛顿方法的变体,如BFGS和DFP算法,这些算法通过近似计算Hessian矩阵,显著降低了计算复杂度。 优化理论部分是本章的另一核心。从无约束优化开始,详细分析了最速下降法、牛顿法、准牛顿法,并引入了拟牛顿法的收敛性理论。对于约束优化问题,本书深入讲解了拉格朗日乘子法、KKT条件,并详细阐述了序列二次规划(SQP)方法,这是求解大规模非线性约束问题的有效手段。 第四部分:常微分方程(ODE)的数值积分 常微分方程在建模物理、化学和生物系统时无处不在。本章专注于建立和分析ODE的数值解法。 首先,介绍了前向和后向欧拉法,并讨论了它们的稳定域。随后,系统地讲解了龙格-库塔(Runge-Kutta, RK)方法族,特别是经典的四阶RK法(RK4)。在稳定性和精度方面,本书对绝对稳定性和A-稳定性进行了详细的数学推导和几何解释,这是选择隐式方法(如后向欧拉法)的关键依据。对于刚性方程组(Stiff Equations),本书专门辟出章节,讲解了隐式欧拉法和BDF(后向微分公式)方法的构建和应用,强调了如何处理大时间步长下的计算稳定性。 第五部分:偏微分方程(PDE)的数值方法概论 本部分是全书的高级主题之一,为后续更专业的数值方法学习打下基础。本书侧重于介绍求解扩散方程(热传导方程)和波动方程(波动方程)的通用框架,避免陷入某一特定离散化技术的细节。 对有限差分法(FDM)进行了详尽的讨论,包括如何利用泰勒展开构建高阶差分近似,并分析了Von Neumann稳定性分析方法,用于判断时间步长和空间步长的耦合关系。 此外,本书还引入了谱方法的基本思想,展示了傅里叶级数展开在周期性问题求解中的高效性,并简要对比了谱方法与局部离散化方法的特点。 本书特色与优势 1. 理论与实践并重: 每种核心算法后都附有详细的稳定性、收敛性分析和计算复杂度的评估。 2. 清晰的算法结构: 所有关键算法均以伪代码形式清晰呈现,便于读者直接转化为计算机程序实现。 3. 深入的矩阵代数背景: 对矩阵性质、特征值和奇异值分解在数值稳定性中的作用进行了强化讲解。 4. 现代计算视角: 关注了大规模计算中的稀疏性处理和预处理技术,这些是现代高性能计算环境中的必备知识。 本书内容全面、论述严谨,是深入理解现代计算科学和进行复杂工程仿真不可或缺的参考教材。