Questions: Convergence of Iterative Methods

5 questions to test your understanding

Score: 0 / 5
Question 1 Multiple Choice

An engineer runs Jacobi iteration on a diagonally dominant system with condition number 50,000. After 200 iterations, convergence is extremely slow. They conclude the large condition number is preventing convergence. What is wrong with this diagnosis?

AJacobi iteration only applies to symmetric matrices, making diagonal dominance irrelevant
BThe condition number measures sensitivity of the solution to data perturbations, not convergence speed — the spectral radius of the iteration matrix, which may be close to 1, determines convergence rate
C200 iterations is simply insufficient; diagonal dominance guarantees convergence but requires at least 1,000 iterations
DDiagonal dominance guarantees fast convergence for Jacobi, so the slow convergence must be a software error
Question 2 Multiple Choice

For the iterative scheme x^(k+1) = Mx^(k) + c, which condition is BOTH necessary AND sufficient for convergence from any starting point?

AM is symmetric positive definite
BThe spectral radius ρ(M) = max|λ_i| < 1
CA is strictly diagonally dominant
DThe condition number κ(A) < 10
Question 3 True / False

If the spectral radius of an iteration matrix is 0.5, the error after 20 iterations is approximately 10^(−6) of the initial error.

TTrue
FFalse
Question 4 True / False

The spectral radius of the iteration matrix and the condition number of A both measure the same underlying property — how difficult a linear system is to solve.

TTrue
FFalse
Question 5 Short Answer

Why is the spectral radius of the iteration matrix, rather than the condition number of A, the correct tool for predicting whether an iterative method converges?

Think about your answer, then reveal below.