Why Is Gauss-Seidel Method's Convergence Rate Double That of Jacobi's?

  • Context: Graduate 
  • Thread starter Thread starter abrowaqas
  • Start date Start date
  • Tags Tags
    Jacobi
Click For Summary
SUMMARY

The Gauss-Seidel method exhibits a convergence rate that is approximately double that of the Jacobi iterative method under specific conditions. This assertion, while commonly accepted, lacks a universal proof due to its dependency on the matrix ordering. In certain scenarios, the Gauss-Seidel method may converge while the Jacobi method does not, complicating the establishment of a definitive general result. The discussion highlights the need for further exploration of counter-examples to validate or refute the convergence claims.

PREREQUISITES
  • Understanding of iterative methods in numerical analysis
  • Familiarity with matrix theory and properties
  • Knowledge of convergence criteria for iterative algorithms
  • Basic proficiency in mathematical proofs and counter-examples
NEXT STEPS
  • Research the convergence criteria for iterative methods in numerical analysis
  • Study the properties of matrix ordering and its impact on convergence
  • Learn about counter-examples in numerical methods, particularly for Gauss-Seidel and Jacobi methods
  • Explore advanced numerical analysis textbooks for proofs related to convergence rates
USEFUL FOR

Mathematicians, numerical analysts, and students studying iterative methods in numerical analysis who seek to understand the nuances of convergence rates between different algorithms.

abrowaqas
Messages
113
Reaction score
0
How we prove that rate of convergence of gauss-Seidel method is approximately twice that of Jacobi iterative method without doing an example itself ?

What's the general proof of this statement ? I didn't fin in any book ?
Can anyone please help me ?
 
Physics news on Phys.org
Your statement about the convergence speeds is not true in general, though it seems to be widely believed (or at least, it's easy to find assertions that it is true, but without any proof, on the web!).

For example see the first few lines of http://www.mit.edu/~jnt/Papers/J025-89-Jacobi_GS.pdf

The convergence rate of Gauss-Siedel is dependent on the ordering of the matrix, and in some cases it may converge for some orderings but not for others. So proving any general result that GS has better convergence than Jacobi is not going to be easy, even if it is true.

Finding an counter-example might be an easier task.
 
Last edited by a moderator:
Thanks Alphazero . I was also thinking of that .but in many books I have found thy statement . That's why i raise this question.
 

Similar threads

  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 3 ·
Replies
3
Views
2K
Replies
20
Views
5K
  • · Replies 4 ·
Replies
4
Views
24K
  • · Replies 5 ·
Replies
5
Views
6K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 17 ·
Replies
17
Views
2K