Globally and superlinearly convergent inexact Newton-Krylov algorithms for solving nonsmooth equations.
Jinhai ChenLiqun QiPublished in: Numer. Linear Algebra Appl. (2010)
Keyphrases
- newton method
- computational cost
- orders of magnitude
- significant improvement
- combinatorial optimization
- image segmentation
- optimization problems
- interior point
- provably convergent
- globally convergent
- quasi newton
- regularized least squares
- convergence rate
- sensitivity analysis
- theoretical analysis
- data structure
- reinforcement learning