Acceleration of Generalized Conditional Gradient Methods

Long Chen; Zeyi Xu

Home/Archive/arXiv

arXiv · 2610.25473

Acceleration of Generalized Conditional Gradient Methods

Long Chen·Zeyi Xu

View original source ↗Read source document ↗Export metadata ↓

Abstract

We develop an adaptive acceleration of generalized conditional gradient methods for convex objectives with a strongly convex regularizer. A computable primal--dual residual certifies Lyapunov decrease without evaluating the regularizer or any Hessian. Each iteration accepts a balanced accelerated trial when the residual is nonpositive; otherwise it takes a mirror-descent step with doubling backtracking and sets the auxiliary parameter explicitly from the dual decrease and the stopping gap. Under uniform strong-convexity and smoothness bounds on the relevant sublevel set, the method converges globally at a linear rate. Near a nondegenerate minimizer, balanced trials pass above a fixed curvature threshold. The adaptive curvature updates then guarantee a positive fraction of accelerated steps and a local accelerated rate. The method applies directly to entropically regularized Jordan--Kinderlehrer--Otto (JKO) steps whose energies contain nonlocal interaction terms with nondiagonal Hessians. Experiments on JKO steps for granular media and for Poisson--Nernst--Planck systems, in one and three dimensions, show that the method reduces both oracle calls and computing time by a large fraction compared with generalized conditional gradient methods with fixed or adaptive step sizes.

Explore related subjects

math.OC

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Long Chen, Zeyi Xu. 2026-10-18. Acceleration of Generalized Conditional Gradient Methods. https://arxiv.org/abs/2610.25473

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Exploring chordal sparsity in semidefinite programming with sparse plus low-rank data matrices

Semidefinite programming (SDP) problems are challenging to solve because of their high dimensionality. However, solving sparse SDP problems with small tree-width are known to be relatively easier because: (1) they can be decomposed into smaller multi-block SDP problems through chordal conversion; (2) they have low-rank optimal solutions. In this paper, we study more general SDP problems whose coefficient matrices have sparse plus low-rank (SPLR) structure. We develop a unified framework to convert such problems into sparse SDP problems with bounded tree-width. Based on this, we derive rank bounds for SDP problems with SPLR structure, which are tight in the worst case.

math.OC↗

arXiv2026

Bellman Optimality of Average-Reward Robust Markov Decision Processes with a Constant Gain

Learning and optimal control under robust Markov decision processes (MDPs) have received increasing attention, yet most existing theory, algorithms, and applications focus on finite-horizon or discounted models. Long-run average-reward formulations, while natural in many operations research and management contexts, remain underexplored. This is primarily because the dynamic programming foundations are technically challenging and only partially understood, with several fundamental questions remaining open. This paper steps toward a general framework for average-reward robust MDPs by analyzing the constant-gain setting. We study the average-reward robust control problem with possible information asymmetries between the controller and an S-rectangular adversary. Our analysis centers on the constant-gain robust Bellman equation, examining both the existence of solutions and their relationship to the optimal average reward. Specifically, we identify when solutions to the robust Bellman equation characterize the optimal average reward and stationary policies, and we provide one-sided weak communication conditions ensuring solutions' existence. These findings expand the dynamic programming theory for average-reward robust MDPs and lay a foundation for robust dynamic decision making under long-run average criteria in operational environments.

math.OC↗

arXiv2026

Service-Induced Congestion in Memory-Constrained LLM Serving

In large language model (LLM) serving, each request accumulates persistent graphics processing unit (GPU) memory during service as its key-value cache grows with every generated token. Under high concurrency, aggregate memory usage therefore increases endogenously over time: the service process itself creates future capacity pressure. When memory capacity is exceeded, systems evict active requests, discarding cached state and restarting them later, which wastes computation and reduces throughput. We develop a discrete-time dynamical model of memory-constrained LLM inference that captures admission, memory growth, and eviction under continuous batching. In the saturated-input regime, the system admits both eviction-free fixed points and limit cycles with evictions. For homogeneous workloads, we show that the eviction-free equilibrium is unstable and that, except for a Lebesgue-measure-zero exact-capture set, the system converges to a unique worst-case limit cycle that is asymptotically stable outside this exceptional set, with throughput losses as large as 50%. For heterogeneous workloads, we prove a stability criterion in the two-class common-input setting and explain how the survival-polynomial mechanism generalizes to multiple classes and heterogeneous-input lengths. Under an input-dominated scaling regime, coprime decoding lengths stabilize the eviction-free equilibrium, while non-coprime lengths create synchronized modes that drive instability. These results characterize when workload heterogeneity desynchronizes completions and helps stabilize memory-constrained serving. More broadly, we identify service-induced congestion as a structural instability mechanism and derive scheduling design principles for sustaining high throughput.

math.OC↗