Trajectory-restricted Polyak-Łojasiewicz, error bound, and quadratic growth conditions produce linear convergence rates governed by the geometry of visited subsets rather than global conditioning.
Neal Parikh and Stephen Boyd
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
math.OC 3years
2026 3verdicts
UNVERDICTED 3representative citing papers
SALM converges globally to M-stationary points under PLCQ when the nonsmooth term is locally Lipschitz continuous, with a counterexample and numerical evidence on sparse portfolio problems.
A variable-metric non-monotone line search method based on the Fukushima regularized gap function is introduced for mixed variational inequalities and equilibrium problems, with global convergence and R-linear rate proved under strong monotonicity.
citing papers explorer
-
Trajectory-Restricted Optimization Conditions and Geometry-Aware Linear Convergence
Trajectory-restricted Polyak-Łojasiewicz, error bound, and quadratic growth conditions produce linear convergence rates governed by the geometry of visited subsets rather than global conditioning.
-
Convergence of the Safeguarded Augmented Lagrangian Method under the Polyak-Lojasiewicz constraint qualification for Constrained Composite Optimization
SALM converges globally to M-stationary points under PLCQ when the nonsmooth term is locally Lipschitz continuous, with a counterexample and numerical evidence on sparse portfolio problems.
-
A Variable-Metric Non-monotone Line Search Method for Mixed Variational Inequalities and Equilibrium Problems
A variable-metric non-monotone line search method based on the Fukushima regularized gap function is introduced for mixed variational inequalities and equilibrium problems, with global convergence and R-linear rate proved under strong monotonicity.