Exercise · ZW-LINE-SEARCH-002 · ★☆☆ Basic
What is the name of the method that reduces \(\alpha\) by multiplying it by \(\rho\) until the condition holds?
Choose one option.
A Backtracking B Exact Newton C Pure random search D Gradient ascent
Show answer
← Back to the question
Explanation Backtracking starts with a large step and shrinks it geometrically. The other options describe different methods, such as ascending the gradient or using second derivatives .
Line search is a procedure that chooses the step size \(\alpha\) along a descent direction \(p_k\) to reduce the value of the objective function .
\[f(x_k+\alpha p_k)\le f(x_k)+c_1\alpha\nabla f(x_k)^\top p_k\]
More practice problems: line search ZW-LINE-SEARCH-001 What does the Armijo condition require of the step α?ZW-LINE-SEARCH-003 What must the direction p_k satisfy for line search to make sense?ZW-LINE-SEARCH-004 What happens if the step α is too large?ZW-LINE-SEARCH-005 Which parameter controls the minimum slope required by the Armijo condition?ZW-LINE-SEARCH-006 With f(x)=x^2, x_k=1, p_k=-2, c_1=0.5 and rho=0.5, which step does backtracking accept if it starts…ZW-LINE-SEARCH-007 With f(x)=x^2, x_k=1, p_k=-2 and α=tfrac12, what is x_k+1?ZW-LINE-SEARCH-008 With ∇ f(x_k)^top p_k=-4, c_1=0.25 and α=2, what is c_1α∇ f(x_k)^top p_k?ZW-LINE-SEARCH-009 With f(x)=x^2, x_k=2, p_k=-4, c_1=0.25, rho=0.5 and starting at α=1, which step does backtracking a…