Download presentation
Presentation is loading. Please wait.
1
Introduction to Scientific Computing II
Steepest Descent & Conjugate Gradients Dr. Miriam Mehl
2
Steepest Descent – Basic Idea
solution of SLE minimization iterative one-dimensional minima direction of steepest descent?
3
Steepest Descent – Principle
4
Steepest Descent – Principle
5
Steepest Descent – Algorithm
6
Steepest Descent – Algorithm II
7
Steepest Descent – Example
initial error after 1 iteration after 10 iterations
8
Steepest Descent – Example
1/128 1/64 1/32 1/16 h 48,629 11,576 2,744 646 iterations
9
Steepest Descent – Convergence
Poisson with 5-point-stencil like Jacobi
10
Steepest Descent – Convergence
11
Conjugate Gradients – Basic Idea
solution of SLE minimization iterative one-dimensional minima no repeating search directions
12
Steepest Descent – Principle
13
Conjugate Gradients – Principle
14
CG – Algorithm
15
Steepest Descent – Example
initial error after 1 iteration after 10 iterations
16
Conjugate Gradients – Example
initial error after 1 iteration after 10 iterations
17
Conjugate Gradients – Example
322 157 76 35 iterations cg 1/128 1/64 1/32 1/16 h 48,629 11,576 2,744 646 iterations sd 16,129 3,969 961 225 #unknowns
18
CG – Convergence Poisson with 5-point-stencil like SOR
no parameter adjustment
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.