Questions: Order of Convergence

5 questions to test your understanding

Score: 0 / 5
Question 1 Multiple Choice

Newton's method has quadratic convergence and bisection has linear convergence on a problem. Starting from the same initial guess x₀, which method is guaranteed to have a smaller error after the first iteration?

ANewton's method, because quadratic convergence always beats linear convergence
BBisection, because it guarantees halving the interval every step
CNeither — convergence order is asymptotic and says nothing about early iterations far from the root
DThey will have the same error after the first iteration
Question 2 Multiple Choice

A method with order 2 (Newton's) requires evaluating both f and f′ per iteration, while a method with order 1.618 (secant) requires only one function evaluation per iteration. If evaluating f′ is very expensive, which method might deliver more accuracy per unit of computational cost?

ANewton's method, because quadratic convergence always dominates in total iterations needed
BThe secant method, because fewer expensive evaluations per step may compensate for slightly lower convergence order
CThey are equivalent in practice because the secant method's order is close enough to 2
DNewton's method, because it requires fewer total iterations to reach any given tolerance
Question 3 True / False

For a method with quadratic convergence, once the error is around 10⁻⁴, the next iteration's error will be approximately 10⁻⁸.

TTrue
FFalse
Question 4 True / False

If a numerical method has quadratic convergence, it will converge to the solution faster than a linear method starting from any initial guess.

TTrue
FFalse
Question 5 Short Answer

Why is 'order of convergence' not the same as 'efficiency' of a numerical method? Give an example.

Think about your answer, then reveal below.