Communication-Efficient Distributed Online Nonconvex Optimization with Time-Varying Constraints (2505.08592v2)

Abstract: This paper considers distributed online nonconvex optimization with time-varying inequality constraints over a network of agents, where the nonconvex local loss and convex local constraint functions can vary arbitrarily across iterations, and the information of them is privately revealed to each agent at each iteration. For a uniformly jointly strongly connected time-varying directed graph, we propose two distributed bandit online primal--dual algorithm with compressed communication to efficiently utilize communication resources in the one-point and two-point bandit feedback settings, respectively. In nonconvex optimization, finding a globally optimal decision is often NP-hard. As a result, the standard regret metric used in online convex optimization becomes inapplicable. To measure the performance of the proposed algorithms, we use a network regret metric grounded in the first-order optimality condition associated with the variational inequality. We show that the compressed algorithms establish sublinear network regret and cumulative constraint violation bounds. Finally, a simulation example is presented to validate the theoretical results.