Advances in Distributed Optimization and Learning

The field of distributed optimization and learning is witnessing significant advancements, with a focus on developing innovative algorithms and frameworks that can efficiently handle complex problems. Researchers are exploring new approaches to address challenges such as communication efficiency, delayed feedback, and non-convexity. Notably, the development of flexible and robust algorithms, such as variants of the ALADIN method, is improving the convergence and performance of distributed optimization systems. Additionally, novel evaluation metrics, like distributed forgetting-factor regret, are being proposed to assess the performance of online optimization algorithms. The application of these advancements can be seen in various domains, including sensor allocation, active noise control, and large-scale global optimization. Noteworthy papers include: Distributed Consensus Optimization with Consensus ALADIN, which proposes a novel algorithm for solving distributed consensus optimization problems. Capacity-Constrained Online Learning with Delays, which establishes matching upper and lower bounds on achievable regret under novel capacity constraints. ALADIN-$\beta$: A Distributed Optimization Algorithm for Solving MPCC Problems, which introduces a distributed structure-splitting reformulation to efficiently solve Mathematical Programs with Complementarity Constraints.

Sources

Distributed Consensus Optimization with Consensus ALADIN

Mixed-gradients Distributed Filtered Reference Least Mean Square Algorithm -- A Robust Distributed Multichannel Active Noise Control Algorithm

An Approach to Analyze Niche Evolution in XCS Models

Capacity-Constrained Online Learning with Delays: Scheduling Frameworks and Regret Trade-offs

Convergence Theory of Flexible ALADIN for Distributed Optimization

Distributed Forgetting-factor Regret-based Online Optimization over Undirected Connected Networks

ALADIN-$\beta$: A Distributed Optimization Algorithm for Solving MPCC Problems

Distributed observer-based leak detection in pipe flow with nonlinear friction

A Novel Two-Phase Cooperative Co-evolution Framework for Large-Scale Global Optimization with Complex Overlapping

Built with on top of