Projection-Free Gradient Tracking Neurodynamic for Distributed Nonnegative Sparse Optimization
Abstract
This paper proposes a continuous-time, projection-free neurodynamic system based on gradient tracking for distributed nonnegative sparse optimization. Unlike traditional methods, our approach avoids projecting states onto the feasible set. Instead, it employs gradient tracking to estimate average gradients under heterogeneous costs, coupling primal and dual dynamics such that feasibility and sparsity emerge naturally from the equilibrium conditions. The system ensures unique solutions through a locally Lipschitz vector field and guarantees that every equilibrium corresponds to an exact KKT point. We establish rigorous convergence guarantees for convex objectives through three distinct analytical layers. These include a centralized kernel rate, a computable distributed rate derived from Jacobian analysis without restrictive stability assumptions, and a global exponential stability proof based on Lur'e system theory. Our analysis quantitatively separates these layers, revealing that the convergence rate is primarily limited by graph sparsity and remains robust to variations in momentum time scales. Numerical experiments on strongly convex quadratic problems validate the theoretical rates, demonstrate effective active set recovery, and show that the system maintains stable tail convergence rates even under strong local heterogeneity.
Related
- Communication-Efficient Differentially Private Gradient Tracking for Distributed Optimization via Local Updates
- Denoising as Projection: Constrained Optimization with Gradient-Guided Diffusion
- A Bilevel Neurodynamic Approach for Distributed SCAD-Regularized Sparse Least Squares
- Comment on "The polarization hierarchy for polynomial optimization over convex bodies, with applications to nonnegative matrix rank"
- Sharp Convergence of Wasserstein Gradient Flows for Spectrally Nonnegative Interaction Energies
BibTeX
@misc{li2026projectionfree,
title = {{Projection-Free Gradient Tracking Neurodynamic for Distributed Nonnegative Sparse Optimization}},
author = {Juntao Li and Qihang Guo and Deyuan Meng},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.07643},
eprint = {2611.07643},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.07643},
}
RIS
TY - UNPB
TI - Projection-Free Gradient Tracking Neurodynamic for Distributed Nonnegative Sparse Optimization
AU - Juntao Li
AU - Qihang Guo
AU - Deyuan Meng
PY - 2026
DA - 2026/11/05
PB - arXiv
DO - 10.48550/arxiv.2611.07643
UR - https://arxiv.org/abs/2611.07643
AB - This paper proposes a continuous-time, projection-free neurodynamic system based on gradient tracking for distributed nonnegative sparse optimization. Unlike traditional methods, our approach avoids projecting states onto the feasible set. Instead, it employs gradient tracking to estimate average gradients under heterogeneous costs, coupling primal and dual dynamics such that feasibility and sparsity emerge naturally from the equilibrium conditions. The system ensures unique solutions through a locally Lipschitz vector field and guarantees that every equilibrium corresponds to an exact KKT point. We establish rigorous convergence guarantees for convex objectives through three distinct analytical layers. These include a centralized kernel rate, a computable distributed rate derived from Jacobian analysis without restrictive stability assumptions, and a global exponential stability proof based on Lur'e system theory. Our analysis quantitatively separates these layers, revealing that the convergence rate is primarily limited by graph sparsity and remains robust to variations in momentum time scales. Numerical experiments on strongly convex quadratic problems validate the theoretical rates, demonstrate effective active set recovery, and show that the system maintains stable tail convergence rates even under strong local heterogeneity.
ER -