Search papers, labs, and topics across Lattice.
This paper introduces a decentralized optimization algorithm for minimizing a sum of smooth and non-smooth convex functions over a network, where each agent possesses a local loss function. The algorithm adaptively adjusts the stepsize using local backtracking and min-consensus protocols, leveraging a three-operator splitting factorization and a novel BCV preconditioning metric. The method achieves sublinear convergence under convexity and linear convergence under strong convexity with a partly smooth non-smooth component.
Adaptive stepsize tuning via local backtracking and min-consensus unlocks efficient decentralized optimization with provable convergence guarantees.
The paper studies decentralized optimization over networks, where agents minimize a sum of {\it locally} smooth (strongly) convex losses and plus a nonsmooth convex extended value term. We propose decentralized methods wherein agents {\it adaptively} adjust their stepsize via local backtracking procedures coupled with lightweight min-consensus protocols. Our design stems from a three-operator splitting factorization applied to an equivalent reformulation of the problem. The reformulation is endowed with a new BCV preconditioning metric (Bertsekas-O'Connor-Vandenberghe), which enables efficient decentralized implementation and local stepsize adjustments. We establish robust convergence guarantees. Under mere convexity, the proposed methods converge with a sublinear rate. Under strong convexity of the sum-function, and assuming the nonsmooth component is partly smooth, we further prove linear convergence. Numerical experiments corroborate the theory and highlight the effectiveness of the proposed adaptive stepsize strategy.
