Derives accessible O(κΦ ln(κΦ ||w*||/ε)) iteration bound for rPDHG on unique-optima LPs, with computable Φ, two-stage performance, and equivalence to stability and sharpness.
Restarted primal-dual hybrid conjugate gradient method for large-scale quadratic programming.arXiv preprint arXiv:2405.16160, 2024
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
math.OC 3verdicts
UNVERDICTED 3representative citing papers
DSPDHG extends PDHG and SPDHG with doubly stochastic block updates and proves O(1/K) ergodic convergence for the expected restricted primal-dual gap plus linear convergence for a restarted variant under quadratic growth.
D-PDLP is the first distributed multi-GPU framework for PDLP that uses 2D grid partitioning of the constraint matrix plus nonzero-aware and random-permutation strategies to scale PDHG iterations with low overhead and full FP64 accuracy.
citing papers explorer
-
Accessible Complexity Bounds for Restarted PDHG on Linear Programs with a Unique Optimizer
Derives accessible O(κΦ ln(κΦ ||w*||/ε)) iteration bound for rPDHG on unique-optima LPs, with computable Φ, two-stage performance, and equivalence to stability and sharpness.
-
On the convergence of doubly stochastic Primal-Dual Hybrid Gradient Method
DSPDHG extends PDHG and SPDHG with doubly stochastic block updates and proves O(1/K) ergodic convergence for the expected restricted primal-dual gap plus linear convergence for a restarted variant under quadratic growth.
-
D-PDLP: Scaling PDLP to Distributed Multi-GPU Systems
D-PDLP is the first distributed multi-GPU framework for PDLP that uses 2D grid partitioning of the constraint matrix plus nonzero-aware and random-permutation strategies to scale PDHG iterations with low overhead and full FP64 accuracy.