Introduction. Math 1080: Numerical Linear Algebra Chapter 4, Iterative Methods. Example: First Order Richardson. Strategy

Size: px
Start display at page:

Download "Introduction. Math 1080: Numerical Linear Algebra Chapter 4, Iterative Methods. Example: First Order Richardson. Strategy"

Transcription

1 Introduction Math 1080: Numerical Linear Algebra Chapter 4, Iterative Methods M. M. Sussman Office Hours: MW 1:45PM-2:45PM, Thack 622 Solve system Ax = b by repeatedly computing residuals r = b Ax Manageable computer memory requirements Store only nonzero entries A Can be implemented on large-scale parallel computers Many different methods available Methods of choice for many problems March / 70 3 / 70 Strategy Computing the residual r = b A x for x an approximate solution is cheap in both operations and storage. Algorithm (Basic Iterative Method) Given an approximate solution x and a maximum number of steps itmax: Compute residual: r = b A x for i = 1:itmax Use r to improve x Compute residual using improved x: r = b A x Use residual and update to estimate accuracy if accuracy is acceptable, exit with converged solution Signal failure if accuracy is not acceptable. Example: First Order Richardson Pick the number ρ > 0 Rewrite Ax = b as ρ(x x) = b Ax, Write ρ(x n+1 x n ) = b Ax n Solve for x n+1 x n+1 = [I 1 ρ A]x n + 1 ρ b. Guess x 0 and iterate using x n+1 = [I 1 ρ A]x n + 1 ρ b. 4 / 70 5 / 70

2 Algorithm (FOR = First Order Richardson) Given ρ > 0, target accuracy tol, maximum number of steps itmax and initial guess x 0 : Compute residual: r 0 = b Ax 0 for n = 1:itmax Compute update n = (1/ρ)r n Compute next approximation x n+1 = x n + n Compute residual r n+1 = b Ax n+1 Estimate residual accuracy criterion r n+1 / b <tol Estimate update accuracy criterion n / x n+1 <tol if both residual and update are acceptable exit with converged solution Signal failure if accuracy is not acceptable. Recall the MPP in 3D u ijk = u(x i, y j, z k ) and f ijk = f (x i, y j, z k ) For a typical point (x i, y j, z k ) Ω, the equation becomes 6u ijk u i+1jk u i 1jk u ij+1k u ij 1k u ijk+1 u ijk 1 = h 2 f ijk, If (x i, y j, z k ) lies on the boundary, u ijk = 0 Take i, j, k = 1,..., N + 1 Boundary values at i = 1 or i = N + 1, j = 1 or j = N + 1, k = 1 or k = N + 1 Choose ρ = 6 and FOR turns into the Jacobi method. 6 / 70 7 / 70 Jacobi iteration in 3d Given: 1. A tolerance tol, 2. A maximum number of iterations itmax 3. Arrays uold, unew and f, each of size (N+1,N+1,N+1) 4. Boundary values of uold and unew filled with zeros for homogeneous Dirichlet b.c. h=1/n for it=1:itmax % initialize solution, delta, residual and rhs norms delta=0 unorm=0 bnorm=0 Jacobi iteration cont d for i=2:n for j=2:n for k=2:n % compute increment au=-( uold(i+1,j,k) + uold(i,j+1,k)... + uold(i,j,k+1) + uold(i-1,j,k)... + uold(i,j-1,k) + uold(i,j,k-1) ) unew(i,j,k)=(h^2*f(i,j,k) - au)/6 % add next term to norms delta=delta + (unew(i,j,k) - uold(i,j,k))^2 unorm=unorm + (unew(i,j,k))^2 bnorm=bnorm + (h^2*f(i,j,k))^2 uold=unew % set uold for next iteration 8 / 70 9 / 70

3 Jacobi iteration norms and residual Jacobi iteration convergence test Remarks % complete norm calculation delta=sqrt(delta) unorm=sqrt(unorm) bnorm=sqrt(bnorm) % compute residual resid=0; for i=2:n for j=2:n for k=2:n au= - ( unew(i+1,j,k) + unew(i,j+1,k)... + unew(i,j,k+1) + unew(i-1,j,k)... + unew(i,j-1,k) + unew(i,j,k-1) ); resid=resid + (h^2*f(i,j,k) - au *unew(i,j,k))^2; resid=sqrt(resid) 10 / 70 % test for convergence if resid <= tol*bnorm & delta <= tol*unorm solution converged return error( convergence failed ) Example: FOR for 1d MPP 11 / 70 In the file jacobi3d.m among the supplemental files on my web site If this algorithm were written to be executed on a computer, the calculation of bnorm would be done once, before the loop began. Two computations of au are really the same and should be combined. Requires 3 (N + 1) (N + 1) (N + 1) arrays Matrix is not stored Not fast enough for prime time u = f (x) = x, 0 < x < 1 u(0) = 0 = u(1) 5 with h = 1 5 leads to the 4 4 tridiagonal linear system whose true solution is 2u 1 u 2 = 1 u 1 +2u 2 u 3 = 2 u 2 +2u 3 u 4 = 3 u 3 +2u 4 = 4. u 1 = 4, u 2 = 7, u 3 = 8, u 4 = / / 70

4 FOR/Jacobi iteration Taking ρ = 2 u1 NEW = uold 2 u2 NEW = uold 1 + u3 OLD 2 u3 NEW = uold 2 + u4 OLD 2 4 = uold 3 2. u NEW Starting from u OLD = u 0 = (0, 0, 0, 0) t u 1 = 1/2 1 3/2 2, u10 = , u20 = steps for 3-digit accuracy for a 4 4 system Homework, u35 = / 70 Matlab example % Jacobi iteration for -u =x/5, u(0)=u(1)=0, h=1/5 utrue=[4; 7; 8; 6]; unew =[0; 0; 0; 0]; for k=1:35 uold = unew; % column vector % equations from slides unew(1) = * uold(2); unew(2) = * (uold(1) + uold(3)); unew(3) = * (uold(2) + uold(4)); unew(4) = * uold(3); [unew, utrue] % print iterates and true solution n=input( ); % wait for keypress [unew, utrue] Preconditioning FOR 15 / 70 Text, Exercises 197, 199, 200. Do 197 and 199 by hand do not try to write a program to do them. In Exercise 200, you should modify jacobi3d.m to handle the 2D case. Do not forget to change ρ to 4 from 6. Given: 1. A N N matrix A 2. A N N matrix M called a Preconditioner 3. A right side vector b 4. An initial guess x 0 n=0 while convergence is not satisfied Obtain x n+1 as the solution of M(x n+1 x n ) = b Ax n n=n+1 16 / / 70

5 Preconditioner Definition: stationary iterative method Pure FOR has M = (1/ρ)I Choose M = A, get solution in 1 iteration Tradeoff between complexity of M and number of iterations An iterative method that can be expressed in the form x n+1 = Bx n + c where B and c do not dep on n is called Stationary. Example Both FOR and preconditioned FOR are stationary. 19 / / 70 Fixed points Residual-Update Form: For a function Φ, a fixed point of Φ(x) is any x satisfying x = Φ(x) A fixed point iteration or Picard iteration is an algorithm approximating x by 1. Guess x 0 2. Repeat x n+1 = Φ(x n ) until convergence. Given x n Repeat until convergence: 1. Compute residual: r n = b Ax n 2. Compute update: n = M 1 r n 3. Perform update: x n+1 = x n + n This is often the way the methods are programmed. 22 / / 70

6 Fixed Point Iteration Form Regular Splitting Form: Stationary iterative method Rewrite as fixed point iteration: Define T = I M 1 A T is the iteration operator. x n+1 = M 1 b + Tx n =: Φ(x n ). This is the form used to analyze convergence and rates of convergence. Rewrite A = M N so N = M A Write Ax = b as Mx = b + Nx Regular Splitting form: Mx n+1 = b + Nx n. Example: FOR M = ρi and N = ρi A so the regular splitting form becomes (ρi)x n+1 = b + (ρi A)x n. 25 / / 70 Jacobi method Theorem 205 Suppose A = D L U 1. D is the diagonal part of A 2. L is the lower triangular part of A 3. U is the upper triangular part of A Jacobi method is given as Dx n+1 = b + (L + U)x n. When D = ρi, The Jacobi method for MPP agrees with FOR. Theorem Consider first order Richardson: T = T FOR = I ρ 1 A Error: e n = x x n Residual: r n = b Ax n Update: n = x n+1 x n all satisfy the same iteration e n+1 = Te n r n+1 = Tr n n+1 = T n 27 / / 70

7 Proof e n+1 = Te n Proof n+1 = T n Since x = ρ 1 b + Tx and x n+1 = ρ 1 b + Tx n, subtraction gives (x x n+1 ) = T (x x n ) and e n+1 = Te n. Start with Subtraction gives So that x n+1 = ρ 1 b + Tx n x n = ρ 1 b + Tx n 1. (x n+1 x n ) = T (x n x n 1 ) n+1 = T n. 29 / / 70 Proof r n+1 = Tr n What the theorem means Since ρx n+1 = ρx n + b Ax n = ρx n + r n multiply by A and add ρb: ρ ( b Ax n+1) = ρ (b Ax n ) Ar n ρr n+1 = ρr n Ar n r n+1 = (I ρ 1 A)r n = Tr n. For methods more complicated than FOR, this part of the proof gives r n+1 = ATA 1 r n and the typical result is n, r n and e n 0 r n and e n can be of widely different sizes at the same rate. If the residuals improve by k significant digits over initial, the errors also improve similarly over the initial error Observable residual behavior indicates error behavior e n+1 = Te n, n+1 = T n, and r n+1 = ATA 1 r n. 31 / / 70

8 Three stopping criteria Another thing to monitor 1. Too Many Iterations: If n itmax stop, and signal failure. 2. Small Residual: Given tol1, check r n b tol1. 3. Small Update: Given tol2, check Monitor α n := r n+1 r n or n+1 n. α < 1 or > 1 suggests convergence or divergence n x n tol2. 4. Stop and signal success if 2 and 3 are satisfied. 33 / / 70 If α is roughly constant Suppose Clearly And, in the limit, so that and hence x N x α n = n+1 n α < 1 x N = x 0 + x = x 0 + x N x = n=n+1 N n=1 n=1 n=n+1 n n n n N+1 1 α The good and the bad Iterative method require minimal storage Need stopping criteria Usually need good preconditioner Need to choose fastest of many available methods 35 / / 70

9 FOR convergence in 1D Investigating convergence Given a, b, for what values of ρ will FOR converge? x n+1 = x n + 1 ρ (b ax n ) Solution x = b/a, and b = ax Subtract x from both sides e n+1 = e n (1/ρ)(ax ax n ) = (1 (a/ρ))e n = te n Hence e n = t n x 0 Convergence when t < 1 If a > 0, 0 < a/2 < ρ If a < 0, ρ < a/2 < 0 Fastest convergence if t = 0, ρ = a Definition 209 Spectral radius of matrix T, spr(t), is the size of the largest eigenvalue spr(t ) = max{ λ : λ = λ(t )}. Theorem 207 The stationary iteration e n+1 = Te n converges for all initial guesses if and only if there is some matrix norm with T < 1. Theorem 208 The stationary iteration e n+1 = Te n converges for all initial guesses if and only if spr(t ) < / / 70 Proofs What about nonsymmetric? If T = ρ < 1, then e n T n e 0 0 If any λ 1, set e 0 to be the associated eigenvector, and convergence fails. If T is symmetric, then there is a basis {v i } consisting of eigenvectors, so that e 0 = i a iv i Hence e n = i a iλ n i v i If all λ < 1, e n 0 % nonsymmetricexample.m A=[ ]; A^2 A^3 for k=1:1000 n(k)=norm(a^k,inf); plot(n) 40 / / 70

10 Similar matrices have same eigenvalues Functions of matrices Definition 212 Matrices B and PBP 1 are similar. Lemma 213 Similar matrices have the same eigenvalues Proof Bφ = λφ PB(P 1 P)φ = λpφ (PBP 1 )ψ = λψ where ψ = Pφ If a function has a series representation, can plug in a square matrix! f (x) = 1 f (A) = A 1 x f (x) = 1 f (A) = I + A x f (x) = e x = 1 + x + x 2 f (x) = x 2 1 f (A) = A 2 I. 2! +... f (A) = ea = n=0 A n n!, eigenvalues of f (A) are f (eigenvalues of A) 42 / 70 Homework 43 / 70 Spectral mapping theorem Let f : C C be an analytic function. If (λ, φ) is an eigenpair for A then (f (λ), φ) is an eigenpair for f (A). Text, 215, 217. ( SMT means Spectral Mapping Theorem ) 44 / / 70

11 Convergence of FOR Theorem 218 Suppose A is SPD. Then FOR converges for any initial guess x 0 provided ρ > λ max (A)/2. Proof Rewrite Ax = b as ρ(x x) = b Ax. Hence the error equation ρ(e n+1 e n ) = Ae n, e n = x x n, Convergence of FOR cont d We know e n 0 provided λ(t ) < 1, or Since λ(a) and ρ are positive. 1 < 1 λ(a)/ρ < < 1 λ max ρ or e n+1 = Te n, T = (I ρ 1 A). so that ρ < ρ λ max Need spr(t ) < 1 f (x) = 1 x/ρ, T = f (A), so λ(t ) = 1 λ(a)/ρ Since A is SPD, λ(a) is real and positive: or ρ > λ max(a). 2 0 < λ min (A) λ(a) λ max (A) <. 47 / / 70 Optimizing ρ spr(t ) = 1 λ/ρ for 0 < λ min λ(a) λ max The smaller T 2 = spr(t ) the faster e n 0. If A = λi, spr(t ) = 1 λ/ρ Can minimize with ρ = λ spr(t ) If 0 < λ min λ(a) λ max ρ 50 / / 70

12 Finding optimal ρ Practical choices of ρ For a given ρ T 2 = spr(t ) = max λ(t ) = max 1 λ(a)/ρ λ(a) Since λ min λ(a) λ max, T 2 = max{ 1 λ min /ρ, 1 λ max /ρ }. Already know T 2 < 1 for ρ > λ max /2 Optimal value ρ is intersection (1 λ min /ρ) = 1 λ max /ρ Easy to estimate λ max as A for any norm Harder to estimate λ min Recall condition number= κ = λ max /λ min ρ = λ max, T 2 = 1 too small! If ρ = λ max (A), T 2 = 1 1/κ If ρ = (λ max + λ min )/2, T 2 = 1 2/(κ + 1). Because of slopes, better slightly large than slightly small Possible to adaptively estimate optimal ρ. 2ρ = λ max + λ min ρ = λ max 2 + λ min 2 52 / / 70 How many iterations? FOR error reduction summary Since e n = Te n 1, e n = T n e 0 If want to reduce initial error by a factor 0.1, e n e 0 T n 0.1 Taking logs and solving: n ln(10)/ ln( T ) Let T = 1 α, for small α ln(1 α) α + O(α 2 ), 1. ρ = λ max (A), FOR requires n ln(10) κ(a) iterations per significant digit of accuracy. 2. ρ = (λ max (A) + λ min (A))/2, FOR requires n ln(10) κ(a)/2 iterations. 3. For MPP, κ(a) = O(h 2 ), FOR requires n = O(h 2 ) iterations per significant digit of accuracy. 4. Much too slow! α = 1/κ for ρ = λ max n ln(10)κ 54 / / 70

13 Homework Gauß-Seidel in 1D MPP Jacobi (FOR with ρ = 2) Gauß-Seidel Exercise E: Matlab investigations of convergence of FOR. h=1/n for it = 1:itmax for i = 2:N unew(i)=h^2*f(i) +... (uold(i-1)+uold(i+1))/2 if convergence satisfied exit uold=unew h=1/n for it = 1:itmax for i = 2:N u(i)=h^2*f(i) +... (u(i-1) + u(i+1))/2 if convergence satisfied exit Algebraic description 56 / 70 Convergence rate 58 / 70 Write A = D + L + U Jacobi: Dx n+1 = b (L + U)x n equivalently D(x n+1 x n ) = b Ax n Gauß-Seidel: (D + U)x n+1 = b Lx n equivalently (D + U)(x n+1 x n ) = b Ax n General form: M(x n+1 x n ) = b Ax n Gauß-Seidel usually takes half as many iterations as Jacobi For the MPP, half of O(h 2 ) is still O(h 2 ) Would like to reduce the exponent! Ingenious choice of M can help, but is often problem-depent. 59 / / 70

14 Relaxation The good and the bad Given ω > 0, a maximum number of iterations itmax and x 0 : for n=1:itmax Compute x n+1 temp by some iterative method Relax x n+1 = ωx n+1 n temp + (1 ω)x if x n+1 is acceptable, exit Good Easy to put into any iterative method program. Good The right choice of ω can reduce the number of iterations Bad The wrong choice of ω can increase the number of iterations When programming, x n+1 temp and x n+1 can take the same storage. Underrelaxation example 62 / 70 Overrelaxation example 63 / 70 Consider e n = ( 0.9)e n Underrelaxation with omega = Final number is e-12 Consider e n = 0.9 e n Overrelaxation with ω = final number = e / / 70

15 Homework SOR for 2D MPP Given an array u of size N+1 by N+1 with boundary values filled with zeros, a maximum number of iterations itmax, a tolerance tol, and an estimate for the optimal omega =omega: Text, Exercise 225, 226. h=1/n for it=1:itmax for i=2:n for j=2:n uold=u(i,j) u(i,j)=h^2*f(i,j)... + (u(i+1,j)+u(i-1,j)+u(i,j+1)+u(i,j-1))/4 u(i,j)=omega*u(i,j)+(1-omega)*uold(i,j) if convergence is satisfied, exit Theorem / 70 How to choose ω optimal? 67 / 70 Theorem (Convergence of SOR) Let A be SPD and let T Jacobi = D 1 (L + U) be the iteration matrix for Jacobi (not SOR). If spr(t Jacobi ) < 1, then SOR converges for any ω with 0 < ω < 2 and there is an optimal choice of ω, known as ω optimal, given by 2 ω optimal = (spr(t Jacobi )) 2 For ω = ω optimal and T SOR, T GaussSeidel the iteration matrices for SOR and Gauss-Seidel respectively, we have spr(t SOR ) = ω optimal 1 < (spr(t GaussSeidel )) 2 spr(t Jacobi ) < 1. SOR convergence rate is the Gauß-Seidel rate squared! From O(h 2 ) to O(h). Known for MPP with simple geometry Can be estimated dynamically 68 / / 70

16 Homework Text, / 70

Math 1080: Numerical Linear Algebra Chapter 4, Iterative Methods

Math 1080: Numerical Linear Algebra Chapter 4, Iterative Methods Math 1080: Numerical Linear Algebra Chapter 4, Iterative Methods M. M. Sussman sussmanm@math.pitt.edu Office Hours: MW 1:45PM-2:45PM, Thack 622 March 2015 1 / 70 Topics Introduction to Iterative Methods

More information

Some definitions. Math 1080: Numerical Linear Algebra Chapter 5, Solving Ax = b by Optimization. A-inner product. Important facts

Some definitions. Math 1080: Numerical Linear Algebra Chapter 5, Solving Ax = b by Optimization. A-inner product. Important facts Some definitions Math 1080: Numerical Linear Algebra Chapter 5, Solving Ax = b by Optimization M. M. Sussman sussmanm@math.pitt.edu Office Hours: MW 1:45PM-2:45PM, Thack 622 A matrix A is SPD (Symmetric

More information

Bindel, Fall 2016 Matrix Computations (CS 6210) Notes for

Bindel, Fall 2016 Matrix Computations (CS 6210) Notes for 1 Iteration basics Notes for 2016-11-07 An iterative solver for Ax = b is produces a sequence of approximations x (k) x. We always stop after finitely many steps, based on some convergence criterion, e.g.

More information

6. Iterative Methods for Linear Systems. The stepwise approach to the solution...

6. Iterative Methods for Linear Systems. The stepwise approach to the solution... 6 Iterative Methods for Linear Systems The stepwise approach to the solution Miriam Mehl: 6 Iterative Methods for Linear Systems The stepwise approach to the solution, January 18, 2013 1 61 Large Sparse

More information

Background. Background. C. T. Kelley NC State University tim C. T. Kelley Background NCSU, Spring / 58

Background. Background. C. T. Kelley NC State University tim C. T. Kelley Background NCSU, Spring / 58 Background C. T. Kelley NC State University tim kelley@ncsu.edu C. T. Kelley Background NCSU, Spring 2012 1 / 58 Notation vectors, matrices, norms l 1 : max col sum... spectral radius scaled integral norms

More information

CAAM 454/554: Stationary Iterative Methods

CAAM 454/554: Stationary Iterative Methods CAAM 454/554: Stationary Iterative Methods Yin Zhang (draft) CAAM, Rice University, Houston, TX 77005 2007, Revised 2010 Abstract Stationary iterative methods for solving systems of linear equations are

More information

Lecture 18 Classical Iterative Methods

Lecture 18 Classical Iterative Methods Lecture 18 Classical Iterative Methods MIT 18.335J / 6.337J Introduction to Numerical Methods Per-Olof Persson November 14, 2006 1 Iterative Methods for Linear Systems Direct methods for solving Ax = b,

More information

COURSE Iterative methods for solving linear systems

COURSE Iterative methods for solving linear systems COURSE 0 4.3. Iterative methods for solving linear systems Because of round-off errors, direct methods become less efficient than iterative methods for large systems (>00 000 variables). An iterative scheme

More information

Chapter 7 Iterative Techniques in Matrix Algebra

Chapter 7 Iterative Techniques in Matrix Algebra Chapter 7 Iterative Techniques in Matrix Algebra Per-Olof Persson persson@berkeley.edu Department of Mathematics University of California, Berkeley Math 128B Numerical Analysis Vector Norms Definition

More information

9. Iterative Methods for Large Linear Systems

9. Iterative Methods for Large Linear Systems EE507 - Computational Techniques for EE Jitkomut Songsiri 9. Iterative Methods for Large Linear Systems introduction splitting method Jacobi method Gauss-Seidel method successive overrelaxation (SOR) 9-1

More information

JACOBI S ITERATION METHOD

JACOBI S ITERATION METHOD ITERATION METHODS These are methods which compute a sequence of progressively accurate iterates to approximate the solution of Ax = b. We need such methods for solving many large linear systems. Sometimes

More information

CLASSICAL ITERATIVE METHODS

CLASSICAL ITERATIVE METHODS CLASSICAL ITERATIVE METHODS LONG CHEN In this notes we discuss classic iterative methods on solving the linear operator equation (1) Au = f, posed on a finite dimensional Hilbert space V = R N equipped

More information

Parallel Numerics, WT 2016/ Iterative Methods for Sparse Linear Systems of Equations. page 1 of 1

Parallel Numerics, WT 2016/ Iterative Methods for Sparse Linear Systems of Equations. page 1 of 1 Parallel Numerics, WT 2016/2017 5 Iterative Methods for Sparse Linear Systems of Equations page 1 of 1 Contents 1 Introduction 1.1 Computer Science Aspects 1.2 Numerical Problems 1.3 Graphs 1.4 Loop Manipulations

More information

Iterative Methods for Solving A x = b

Iterative Methods for Solving A x = b Iterative Methods for Solving A x = b A good (free) online source for iterative methods for solving A x = b is given in the description of a set of iterative solvers called templates found at netlib: http

More information

Notes for CS542G (Iterative Solvers for Linear Systems)

Notes for CS542G (Iterative Solvers for Linear Systems) Notes for CS542G (Iterative Solvers for Linear Systems) Robert Bridson November 20, 2007 1 The Basics We re now looking at efficient ways to solve the linear system of equations Ax = b where in this course,

More information

Classical iterative methods for linear systems

Classical iterative methods for linear systems Classical iterative methods for linear systems Ed Bueler MATH 615 Numerical Analysis of Differential Equations 27 February 1 March, 2017 Ed Bueler (MATH 615 NADEs) Classical iterative methods for linear

More information

NUMERICAL ALGORITHMS FOR A SECOND ORDER ELLIPTIC BVP

NUMERICAL ALGORITHMS FOR A SECOND ORDER ELLIPTIC BVP ANALELE ŞTIINŢIFICE ALE UNIVERSITĂŢII AL.I. CUZA DIN IAŞI (S.N. MATEMATICĂ, Tomul LIII, 2007, f.1 NUMERICAL ALGORITHMS FOR A SECOND ORDER ELLIPTIC BVP BY GINA DURA and RĂZVAN ŞTEFĂNESCU Abstract. The aim

More information

Math 471 (Numerical methods) Chapter 3 (second half). System of equations

Math 471 (Numerical methods) Chapter 3 (second half). System of equations Math 47 (Numerical methods) Chapter 3 (second half). System of equations Overlap 3.5 3.8 of Bradie 3.5 LU factorization w/o pivoting. Motivation: ( ) A I Gaussian Elimination (U L ) where U is upper triangular

More information

Iterative techniques in matrix algebra

Iterative techniques in matrix algebra Iterative techniques in matrix algebra Tsung-Ming Huang Department of Mathematics National Taiwan Normal University, Taiwan September 12, 2015 Outline 1 Norms of vectors and matrices 2 Eigenvalues and

More information

The amount of work to construct each new guess from the previous one should be a small multiple of the number of nonzeros in A.

The amount of work to construct each new guess from the previous one should be a small multiple of the number of nonzeros in A. AMSC/CMSC 661 Scientific Computing II Spring 2005 Solution of Sparse Linear Systems Part 2: Iterative methods Dianne P. O Leary c 2005 Solving Sparse Linear Systems: Iterative methods The plan: Iterative

More information

Iterative Methods. Splitting Methods

Iterative Methods. Splitting Methods Iterative Methods Splitting Methods 1 Direct Methods Solving Ax = b using direct methods. Gaussian elimination (using LU decomposition) Variants of LU, including Crout and Doolittle Other decomposition

More information

Jordan Journal of Mathematics and Statistics (JJMS) 5(3), 2012, pp A NEW ITERATIVE METHOD FOR SOLVING LINEAR SYSTEMS OF EQUATIONS

Jordan Journal of Mathematics and Statistics (JJMS) 5(3), 2012, pp A NEW ITERATIVE METHOD FOR SOLVING LINEAR SYSTEMS OF EQUATIONS Jordan Journal of Mathematics and Statistics JJMS) 53), 2012, pp.169-184 A NEW ITERATIVE METHOD FOR SOLVING LINEAR SYSTEMS OF EQUATIONS ADEL H. AL-RABTAH Abstract. The Jacobi and Gauss-Seidel iterative

More information

Here is an example of a block diagonal matrix with Jordan Blocks on the diagonal: J

Here is an example of a block diagonal matrix with Jordan Blocks on the diagonal: J Class Notes 4: THE SPECTRAL RADIUS, NORM CONVERGENCE AND SOR. Math 639d Due Date: Feb. 7 (updated: February 5, 2018) In the first part of this week s reading, we will prove Theorem 2 of the previous class.

More information

Iterative Methods for Ax=b

Iterative Methods for Ax=b 1 FUNDAMENTALS 1 Iterative Methods for Ax=b 1 Fundamentals consider the solution of the set of simultaneous equations Ax = b where A is a square matrix, n n and b is a right hand vector. We write the iterative

More information

Math/Phys/Engr 428, Math 529/Phys 528 Numerical Methods - Summer Homework 3 Due: Tuesday, July 3, 2018

Math/Phys/Engr 428, Math 529/Phys 528 Numerical Methods - Summer Homework 3 Due: Tuesday, July 3, 2018 Math/Phys/Engr 428, Math 529/Phys 528 Numerical Methods - Summer 28. (Vector and Matrix Norms) Homework 3 Due: Tuesday, July 3, 28 Show that the l vector norm satisfies the three properties (a) x for x

More information

Math 5630: Iterative Methods for Systems of Equations Hung Phan, UMass Lowell March 22, 2018

Math 5630: Iterative Methods for Systems of Equations Hung Phan, UMass Lowell March 22, 2018 1 Linear Systems Math 5630: Iterative Methods for Systems of Equations Hung Phan, UMass Lowell March, 018 Consider the system 4x y + z = 7 4x 8y + z = 1 x + y + 5z = 15. We then obtain x = 1 4 (7 + y z)

More information

Computational Linear Algebra

Computational Linear Algebra Computational Linear Algebra PD Dr. rer. nat. habil. Ralf Peter Mundani Computation in Engineering / BGU Scientific Computing in Computer Science / INF Winter Term 2017/18 Part 3: Iterative Methods PD

More information

EXAMPLES OF CLASSICAL ITERATIVE METHODS

EXAMPLES OF CLASSICAL ITERATIVE METHODS EXAMPLES OF CLASSICAL ITERATIVE METHODS In these lecture notes we revisit a few classical fixpoint iterations for the solution of the linear systems of equations. We focus on the algebraic and algorithmic

More information

Iterative Solution methods

Iterative Solution methods p. 1/28 TDB NLA Parallel Algorithms for Scientific Computing Iterative Solution methods p. 2/28 TDB NLA Parallel Algorithms for Scientific Computing Basic Iterative Solution methods The ideas to use iterative

More information

Numerical Methods I Non-Square and Sparse Linear Systems

Numerical Methods I Non-Square and Sparse Linear Systems Numerical Methods I Non-Square and Sparse Linear Systems Aleksandar Donev Courant Institute, NYU 1 donev@courant.nyu.edu 1 MATH-GA 2011.003 / CSCI-GA 2945.003, Fall 2014 September 25th, 2014 A. Donev (Courant

More information

Conjugate gradient method. Descent method. Conjugate search direction. Conjugate Gradient Algorithm (294)

Conjugate gradient method. Descent method. Conjugate search direction. Conjugate Gradient Algorithm (294) Conjugate gradient method Descent method Hestenes, Stiefel 1952 For A N N SPD In exact arithmetic, solves in N steps In real arithmetic No guaranteed stopping Often converges in many fewer than N steps

More information

Iterative Methods and Multigrid

Iterative Methods and Multigrid Iterative Methods and Multigrid Part 1: Introduction to Multigrid 2000 Eric de Sturler 1 12/02/09 MG01.prz Basic Iterative Methods (1) Nonlinear equation: f(x) = 0 Rewrite as x = F(x), and iterate x i+1

More information

Computational Economics and Finance

Computational Economics and Finance Computational Economics and Finance Part II: Linear Equations Spring 2016 Outline Back Substitution, LU and other decomposi- Direct methods: tions Error analysis and condition numbers Iterative methods:

More information

Math Introduction to Numerical Analysis - Class Notes. Fernando Guevara Vasquez. Version Date: January 17, 2012.

Math Introduction to Numerical Analysis - Class Notes. Fernando Guevara Vasquez. Version Date: January 17, 2012. Math 5620 - Introduction to Numerical Analysis - Class Notes Fernando Guevara Vasquez Version 1990. Date: January 17, 2012. 3 Contents 1. Disclaimer 4 Chapter 1. Iterative methods for solving linear systems

More information

The Conjugate Gradient Method

The Conjugate Gradient Method The Conjugate Gradient Method Classical Iterations We have a problem, We assume that the matrix comes from a discretization of a PDE. The best and most popular model problem is, The matrix will be as large

More information

CHAPTER 5. Basic Iterative Methods

CHAPTER 5. Basic Iterative Methods Basic Iterative Methods CHAPTER 5 Solve Ax = f where A is large and sparse (and nonsingular. Let A be split as A = M N in which M is nonsingular, and solving systems of the form Mz = r is much easier than

More information

Topics. The CG Algorithm Algorithmic Options CG s Two Main Convergence Theorems

Topics. The CG Algorithm Algorithmic Options CG s Two Main Convergence Theorems Topics The CG Algorithm Algorithmic Options CG s Two Main Convergence Theorems What about non-spd systems? Methods requiring small history Methods requiring large history Summary of solvers 1 / 52 Conjugate

More information

PowerPoints organized by Dr. Michael R. Gustafson II, Duke University

PowerPoints organized by Dr. Michael R. Gustafson II, Duke University Part 3 Chapter 10 LU Factorization PowerPoints organized by Dr. Michael R. Gustafson II, Duke University All images copyright The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

More information

Lecture Note 7: Iterative methods for solving linear systems. Xiaoqun Zhang Shanghai Jiao Tong University

Lecture Note 7: Iterative methods for solving linear systems. Xiaoqun Zhang Shanghai Jiao Tong University Lecture Note 7: Iterative methods for solving linear systems Xiaoqun Zhang Shanghai Jiao Tong University Last updated: December 24, 2014 1.1 Review on linear algebra Norms of vectors and matrices vector

More information

Solving Linear Systems

Solving Linear Systems Solving Linear Systems Iterative Solutions Methods Philippe B. Laval KSU Fall 207 Philippe B. Laval (KSU) Linear Systems Fall 207 / 2 Introduction We continue looking how to solve linear systems of the

More information

The Solution of Linear Systems AX = B

The Solution of Linear Systems AX = B Chapter 2 The Solution of Linear Systems AX = B 21 Upper-triangular Linear Systems We will now develop the back-substitution algorithm, which is useful for solving a linear system of equations that has

More information

DEN: Linear algebra numerical view (GEM: Gauss elimination method for reducing a full rank matrix to upper-triangular

DEN: Linear algebra numerical view (GEM: Gauss elimination method for reducing a full rank matrix to upper-triangular form) Given: matrix C = (c i,j ) n,m i,j=1 ODE and num math: Linear algebra (N) [lectures] c phabala 2016 DEN: Linear algebra numerical view (GEM: Gauss elimination method for reducing a full rank matrix

More information

Lab 1: Iterative Methods for Solving Linear Systems

Lab 1: Iterative Methods for Solving Linear Systems Lab 1: Iterative Methods for Solving Linear Systems January 22, 2017 Introduction Many real world applications require the solution to very large and sparse linear systems where direct methods such as

More information

Numerical Methods - Numerical Linear Algebra

Numerical Methods - Numerical Linear Algebra Numerical Methods - Numerical Linear Algebra Y. K. Goh Universiti Tunku Abdul Rahman 2013 Y. K. Goh (UTAR) Numerical Methods - Numerical Linear Algebra I 2013 1 / 62 Outline 1 Motivation 2 Solving Linear

More information

Introduction and Stationary Iterative Methods

Introduction and Stationary Iterative Methods Introduction and C. T. Kelley NC State University tim kelley@ncsu.edu Research Supported by NSF, DOE, ARO, USACE DTU ITMAN, 2011 Outline Notation and Preliminaries General References What you Should Know

More information

4.6 Iterative Solvers for Linear Systems

4.6 Iterative Solvers for Linear Systems 4.6 Iterative Solvers for Linear Systems Why use iterative methods? Virtually all direct methods for solving Ax = b require O(n 3 ) floating point operations. In practical applications the matrix A often

More information

Iterative Methods for Sparse Linear Systems

Iterative Methods for Sparse Linear Systems Iterative Methods for Sparse Linear Systems Luca Bergamaschi e-mail: berga@dmsa.unipd.it - http://www.dmsa.unipd.it/ berga Department of Mathematical Methods and Models for Scientific Applications University

More information

Chapter 2. Solving Systems of Equations. 2.1 Gaussian elimination

Chapter 2. Solving Systems of Equations. 2.1 Gaussian elimination Chapter 2 Solving Systems of Equations A large number of real life applications which are resolved through mathematical modeling will end up taking the form of the following very simple looking matrix

More information

Solving Linear Systems

Solving Linear Systems Solving Linear Systems Iterative Solutions Methods Philippe B. Laval KSU Fall 2015 Philippe B. Laval (KSU) Linear Systems Fall 2015 1 / 12 Introduction We continue looking how to solve linear systems of

More information

Process Model Formulation and Solution, 3E4

Process Model Formulation and Solution, 3E4 Process Model Formulation and Solution, 3E4 Section B: Linear Algebraic Equations Instructor: Kevin Dunn dunnkg@mcmasterca Department of Chemical Engineering Course notes: Dr Benoît Chachuat 06 October

More information

9.1 Preconditioned Krylov Subspace Methods

9.1 Preconditioned Krylov Subspace Methods Chapter 9 PRECONDITIONING 9.1 Preconditioned Krylov Subspace Methods 9.2 Preconditioned Conjugate Gradient 9.3 Preconditioned Generalized Minimal Residual 9.4 Relaxation Method Preconditioners 9.5 Incomplete

More information

Chapter 12: Iterative Methods

Chapter 12: Iterative Methods ES 40: Scientific and Engineering Computation. Uchechukwu Ofoegbu Temple University Chapter : Iterative Methods ES 40: Scientific and Engineering Computation. Gauss-Seidel Method The Gauss-Seidel method

More information

Notes on PCG for Sparse Linear Systems

Notes on PCG for Sparse Linear Systems Notes on PCG for Sparse Linear Systems Luca Bergamaschi Department of Civil Environmental and Architectural Engineering University of Padova e-mail luca.bergamaschi@unipd.it webpage www.dmsa.unipd.it/

More information

Kasetsart University Workshop. Multigrid methods: An introduction

Kasetsart University Workshop. Multigrid methods: An introduction Kasetsart University Workshop Multigrid methods: An introduction Dr. Anand Pardhanani Mathematics Department Earlham College Richmond, Indiana USA pardhan@earlham.edu A copy of these slides is available

More information

Iterative methods for Linear System

Iterative methods for Linear System Iterative methods for Linear System JASS 2009 Student: Rishi Patil Advisor: Prof. Thomas Huckle Outline Basics: Matrices and their properties Eigenvalues, Condition Number Iterative Methods Direct and

More information

7.3 The Jacobi and Gauss-Siedel Iterative Techniques. Problem: To solve Ax = b for A R n n. Methodology: Iteratively approximate solution x. No GEPP.

7.3 The Jacobi and Gauss-Siedel Iterative Techniques. Problem: To solve Ax = b for A R n n. Methodology: Iteratively approximate solution x. No GEPP. 7.3 The Jacobi and Gauss-Siedel Iterative Techniques Problem: To solve Ax = b for A R n n. Methodology: Iteratively approximate solution x. No GEPP. 7.3 The Jacobi and Gauss-Siedel Iterative Techniques

More information

Preliminary/Qualifying Exam in Numerical Analysis (Math 502a) Spring 2012

Preliminary/Qualifying Exam in Numerical Analysis (Math 502a) Spring 2012 Instructions Preliminary/Qualifying Exam in Numerical Analysis (Math 502a) Spring 2012 The exam consists of four problems, each having multiple parts. You should attempt to solve all four problems. 1.

More information

Theory of Iterative Methods

Theory of Iterative Methods Based on Strang s Introduction to Applied Mathematics Theory of Iterative Methods The Iterative Idea To solve Ax = b, write Mx (k+1) = (M A)x (k) + b, k = 0, 1,,... Then the error e (k) x (k) x satisfies

More information

ITERATIVE METHODS BASED ON KRYLOV SUBSPACES

ITERATIVE METHODS BASED ON KRYLOV SUBSPACES ITERATIVE METHODS BASED ON KRYLOV SUBSPACES LONG CHEN We shall present iterative methods for solving linear algebraic equation Au = b based on Krylov subspaces We derive conjugate gradient (CG) method

More information

CME342 Parallel Methods in Numerical Analysis. Matrix Computation: Iterative Methods II. Sparse Matrix-vector Multiplication.

CME342 Parallel Methods in Numerical Analysis. Matrix Computation: Iterative Methods II. Sparse Matrix-vector Multiplication. CME342 Parallel Methods in Numerical Analysis Matrix Computation: Iterative Methods II Outline: CG & its parallelization. Sparse Matrix-vector Multiplication. 1 Basic iterative methods: Ax = b r = b Ax

More information

Review of matrices. Let m, n IN. A rectangle of numbers written like A =

Review of matrices. Let m, n IN. A rectangle of numbers written like A = Review of matrices Let m, n IN. A rectangle of numbers written like a 11 a 12... a 1n a 21 a 22... a 2n A =...... a m1 a m2... a mn where each a ij IR is called a matrix with m rows and n columns or an

More information

Lecture 4 Basic Iterative Methods I

Lecture 4 Basic Iterative Methods I March 26, 2018 Lecture 4 Basic Iterative Methods I A The Power Method Let A be an n n with eigenvalues λ 1,...,λ n counted according to multiplicity. We assume the eigenvalues to be ordered in absolute

More information

AMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 23: GMRES and Other Krylov Subspace Methods; Preconditioning

AMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 23: GMRES and Other Krylov Subspace Methods; Preconditioning AMS526: Numerical Analysis I (Numerical Linear Algebra) Lecture 23: GMRES and Other Krylov Subspace Methods; Preconditioning Xiangmin Jiao SUNY Stony Brook Xiangmin Jiao Numerical Analysis I 1 / 18 Outline

More information

7.2 Steepest Descent and Preconditioning

7.2 Steepest Descent and Preconditioning 7.2 Steepest Descent and Preconditioning Descent methods are a broad class of iterative methods for finding solutions of the linear system Ax = b for symmetric positive definite matrix A R n n. Consider

More information

Next topics: Solving systems of linear equations

Next topics: Solving systems of linear equations Next topics: Solving systems of linear equations 1 Gaussian elimination (today) 2 Gaussian elimination with partial pivoting (Week 9) 3 The method of LU-decomposition (Week 10) 4 Iterative techniques:

More information

Cache Oblivious Stencil Computations

Cache Oblivious Stencil Computations Cache Oblivious Stencil Computations S. HUNOLD J. L. TRÄFF F. VERSACI Lectures on High Performance Computing 13 April 2015 F. Versaci (TU Wien) Cache Oblivious Stencil Computations 13 April 2015 1 / 19

More information

30.5. Iterative Methods for Systems of Equations. Introduction. Prerequisites. Learning Outcomes

30.5. Iterative Methods for Systems of Equations. Introduction. Prerequisites. Learning Outcomes Iterative Methods for Systems of Equations 0.5 Introduction There are occasions when direct methods (like Gaussian elimination or the use of an LU decomposition) are not the best way to solve a system

More information

Dense Matrices for Biofluids Applications

Dense Matrices for Biofluids Applications Dense Matrices for Biofluids Applications by Liwei Chen A Project Report Submitted to the Faculty of the WORCESTER POLYTECHNIC INSTITUTE In partial fulfillment of the requirements for the Degree of Master

More information

Notes on Some Methods for Solving Linear Systems

Notes on Some Methods for Solving Linear Systems Notes on Some Methods for Solving Linear Systems Dianne P. O Leary, 1983 and 1999 and 2007 September 25, 2007 When the matrix A is symmetric and positive definite, we have a whole new class of algorithms

More information

From Stationary Methods to Krylov Subspaces

From Stationary Methods to Krylov Subspaces Week 6: Wednesday, Mar 7 From Stationary Methods to Krylov Subspaces Last time, we discussed stationary methods for the iterative solution of linear systems of equations, which can generally be written

More information

Section 4.1 The Power Method

Section 4.1 The Power Method Section 4.1 The Power Method Key terms Dominant eigenvalue Eigenpair Infinity norm and 2-norm Power Method Scaled Power Method Power Method for symmetric matrices The notation used varies a bit from that

More information

Math 577 Assignment 7

Math 577 Assignment 7 Math 577 Assignment 7 Thanks for Yu Cao 1. Solution. The linear system being solved is Ax = 0, where A is a (n 1 (n 1 matrix such that 2 1 1 2 1 A =......... 1 2 1 1 2 and x = (U 1, U 2,, U n 1. By the

More information

Numerical Programming I (for CSE)

Numerical Programming I (for CSE) Technische Universität München WT 1/13 Fakultät für Mathematik Prof. Dr. M. Mehl B. Gatzhammer January 1, 13 Numerical Programming I (for CSE) Tutorial 1: Iterative Methods 1) Relaxation Methods a) Let

More information

Goal: to construct some general-purpose algorithms for solving systems of linear Equations

Goal: to construct some general-purpose algorithms for solving systems of linear Equations Chapter IV Solving Systems of Linear Equations Goal: to construct some general-purpose algorithms for solving systems of linear Equations 4.6 Solution of Equations by Iterative Methods 4.6 Solution of

More information

MATH 571: Computational Assignment #2

MATH 571: Computational Assignment #2 MATH 571: Computational Assignment #2 Due on Tuesday, November 26, 2013 TTH 12:pm Wenqiang Feng 1 MATH 571 ( TTH 12:pm): Computational Assignment #2 Contents Problem 1 3 Problem 2 8 Page 2 of 9 MATH 571

More information

Iterative Methods for Linear Systems

Iterative Methods for Linear Systems Iterative Methods for Linear Systems 1. Introduction: Direct solvers versus iterative solvers In many applications we have to solve a linear system Ax = b with A R n n and b R n given. If n is large the

More information

Chapter 7. Iterative methods for large sparse linear systems. 7.1 Sparse matrix algebra. Large sparse matrices

Chapter 7. Iterative methods for large sparse linear systems. 7.1 Sparse matrix algebra. Large sparse matrices Chapter 7 Iterative methods for large sparse linear systems In this chapter we revisit the problem of solving linear systems of equations, but now in the context of large sparse systems. The price to pay

More information

Boundary Value Problems - Solving 3-D Finite-Difference problems Jacob White

Boundary Value Problems - Solving 3-D Finite-Difference problems Jacob White Introduction to Simulation - Lecture 2 Boundary Value Problems - Solving 3-D Finite-Difference problems Jacob White Thanks to Deepak Ramaswamy, Michal Rewienski, and Karen Veroy Outline Reminder about

More information

Lecture 11. Fast Linear Solvers: Iterative Methods. J. Chaudhry. Department of Mathematics and Statistics University of New Mexico

Lecture 11. Fast Linear Solvers: Iterative Methods. J. Chaudhry. Department of Mathematics and Statistics University of New Mexico Lecture 11 Fast Linear Solvers: Iterative Methods J. Chaudhry Department of Mathematics and Statistics University of New Mexico J. Chaudhry (UNM) Math/CS 375 1 / 23 Summary: Complexity of Linear Solves

More information

Motivation: Sparse matrices and numerical PDE's

Motivation: Sparse matrices and numerical PDE's Lecture 20: Numerical Linear Algebra #4 Iterative methods and Eigenproblems Outline 1) Motivation: beyond LU for Ax=b A little PDE's and sparse matrices A) Temperature Equation B) Poisson Equation 2) Splitting

More information

ON A HOMOTOPY BASED METHOD FOR SOLVING SYSTEMS OF LINEAR EQUATIONS

ON A HOMOTOPY BASED METHOD FOR SOLVING SYSTEMS OF LINEAR EQUATIONS TWMS J. Pure Appl. Math., V.6, N.1, 2015, pp.15-26 ON A HOMOTOPY BASED METHOD FOR SOLVING SYSTEMS OF LINEAR EQUATIONS J. SAEIDIAN 1, E. BABOLIAN 1, A. AZIZI 2 Abstract. A new iterative method is proposed

More information

THE EIGENVALUE PROBLEM

THE EIGENVALUE PROBLEM THE EIGENVALUE PROBLEM Let A be an n n square matrix. If there is a number λ and a column vector v 0 for which Av = λv then we say λ is an eigenvalue of A and v is an associated eigenvector. Note that

More information

Algebra C Numerical Linear Algebra Sample Exam Problems

Algebra C Numerical Linear Algebra Sample Exam Problems Algebra C Numerical Linear Algebra Sample Exam Problems Notation. Denote by V a finite-dimensional Hilbert space with inner product (, ) and corresponding norm. The abbreviation SPD is used for symmetric

More information

Sparse Linear Systems. Iterative Methods for Sparse Linear Systems. Motivation for Studying Sparse Linear Systems. Partial Differential Equations

Sparse Linear Systems. Iterative Methods for Sparse Linear Systems. Motivation for Studying Sparse Linear Systems. Partial Differential Equations Sparse Linear Systems Iterative Methods for Sparse Linear Systems Matrix Computations and Applications, Lecture C11 Fredrik Bengzon, Robert Söderlund We consider the problem of solving the linear system

More information

LINEAR SYSTEMS (11) Intensive Computation

LINEAR SYSTEMS (11) Intensive Computation LINEAR SYSTEMS () Intensive Computation 27-8 prof. Annalisa Massini Viviana Arrigoni EXACT METHODS:. GAUSSIAN ELIMINATION. 2. CHOLESKY DECOMPOSITION. ITERATIVE METHODS:. JACOBI. 2. GAUSS-SEIDEL 2 CHOLESKY

More information

Stabilization and Acceleration of Algebraic Multigrid Method

Stabilization and Acceleration of Algebraic Multigrid Method Stabilization and Acceleration of Algebraic Multigrid Method Recursive Projection Algorithm A. Jemcov J.P. Maruszewski Fluent Inc. October 24, 2006 Outline 1 Need for Algorithm Stabilization and Acceleration

More information

Computational Methods. Systems of Linear Equations

Computational Methods. Systems of Linear Equations Computational Methods Systems of Linear Equations Manfred Huber 2010 1 Systems of Equations Often a system model contains multiple variables (parameters) and contains multiple equations Multiple equations

More information

SOLVING SPARSE LINEAR SYSTEMS OF EQUATIONS. Chao Yang Computational Research Division Lawrence Berkeley National Laboratory Berkeley, CA, USA

SOLVING SPARSE LINEAR SYSTEMS OF EQUATIONS. Chao Yang Computational Research Division Lawrence Berkeley National Laboratory Berkeley, CA, USA 1 SOLVING SPARSE LINEAR SYSTEMS OF EQUATIONS Chao Yang Computational Research Division Lawrence Berkeley National Laboratory Berkeley, CA, USA 2 OUTLINE Sparse matrix storage format Basic factorization

More information

Poisson Equation in 2D

Poisson Equation in 2D A Parallel Strategy Department of Mathematics and Statistics McMaster University March 31, 2010 Outline Introduction 1 Introduction Motivation Discretization Iterative Methods 2 Additive Schwarz Method

More information

Solving Sparse Linear Systems: Iterative methods

Solving Sparse Linear Systems: Iterative methods Scientific Computing with Case Studies SIAM Press, 2009 http://www.cs.umd.edu/users/oleary/sccs Lecture Notes for Unit VII Sparse Matrix Computations Part 2: Iterative Methods Dianne P. O Leary c 2008,2010

More information

Solving Sparse Linear Systems: Iterative methods

Solving Sparse Linear Systems: Iterative methods Scientific Computing with Case Studies SIAM Press, 2009 http://www.cs.umd.edu/users/oleary/sccswebpage Lecture Notes for Unit VII Sparse Matrix Computations Part 2: Iterative Methods Dianne P. O Leary

More information

FEM and Sparse Linear System Solving

FEM and Sparse Linear System Solving FEM & sparse system solving, Lecture 7, Nov 3, 2017 1/46 Lecture 7, Nov 3, 2015: Introduction to Iterative Solvers: Stationary Methods http://people.inf.ethz.ch/arbenz/fem16 Peter Arbenz Computer Science

More information

Numerical Solution Techniques in Mechanical and Aerospace Engineering

Numerical Solution Techniques in Mechanical and Aerospace Engineering Numerical Solution Techniques in Mechanical and Aerospace Engineering Chunlei Liang LECTURE 3 Solvers of linear algebraic equations 3.1. Outline of Lecture Finite-difference method for a 2D elliptic PDE

More information

Jae Heon Yun and Yu Du Han

Jae Heon Yun and Yu Du Han Bull. Korean Math. Soc. 39 (2002), No. 3, pp. 495 509 MODIFIED INCOMPLETE CHOLESKY FACTORIZATION PRECONDITIONERS FOR A SYMMETRIC POSITIVE DEFINITE MATRIX Jae Heon Yun and Yu Du Han Abstract. We propose

More information

Introduction to Scientific Computing

Introduction to Scientific Computing (Lecture 5: Linear system of equations / Matrix Splitting) Bojana Rosić, Thilo Moshagen Institute of Scientific Computing Motivation Let us resolve the problem scheme by using Kirchhoff s laws: the algebraic

More information

3.2 Iterative Solution Methods for Solving Linear

3.2 Iterative Solution Methods for Solving Linear 22 CHAPTER 3. NUMERICAL LINEAR ALGEBRA 3.2 Iterative Solution Methods for Solving Linear Systems 3.2.1 Introduction We continue looking how to solve linear systems of the form Ax = b where A = (a ij is

More information

HOMEWORK 10 SOLUTIONS

HOMEWORK 10 SOLUTIONS HOMEWORK 10 SOLUTIONS MATH 170A Problem 0.1. Watkins 8.3.10 Solution. The k-th error is e (k) = G k e (0). As discussed before, that means that e (k+j) ρ(g) k, i.e., the norm of the error is approximately

More information

Lecture 16 Methods for System of Linear Equations (Linear Systems) Songting Luo. Department of Mathematics Iowa State University

Lecture 16 Methods for System of Linear Equations (Linear Systems) Songting Luo. Department of Mathematics Iowa State University Lecture 16 Methods for System of Linear Equations (Linear Systems) Songting Luo Department of Mathematics Iowa State University MATH 481 Numerical Methods for Differential Equations Songting Luo ( Department

More information

1. Nonlinear Equations. This lecture note excerpted parts from Michael Heath and Max Gunzburger. f(x) = 0

1. Nonlinear Equations. This lecture note excerpted parts from Michael Heath and Max Gunzburger. f(x) = 0 Numerical Analysis 1 1. Nonlinear Equations This lecture note excerpted parts from Michael Heath and Max Gunzburger. Given function f, we seek value x for which where f : D R n R n is nonlinear. f(x) =

More information

Analysis of Iterative Methods for Solving Sparse Linear Systems C. David Levermore 9 May 2013

Analysis of Iterative Methods for Solving Sparse Linear Systems C. David Levermore 9 May 2013 Analysis of Iterative Methods for Solving Sparse Linear Systems C. David Levermore 9 May 2013 1. General Iterative Methods 1.1. Introduction. Many applications lead to N N linear algebraic systems of the

More information