Agent-Supervisor Coordination for Decentralized Event-Triggered Optimization

TL;DR

This paper introduces decentralized, resource-aware coordination schemes for network optimization that enable supervised agents to asynchronously converge to optimal solutions while maintaining feasibility, demonstrated through power system simulations.

Contribution

It presents a novel feedback-based, asynchronous coordination method for agents to solve network optimization problems with coupling costs, ensuring convergence and feasibility.

Findings

Guarantees asymptotic convergence to the optimizer

Ensures anytime feasibility during operation

Demonstrates effectiveness through power system simulations

Abstract

This paper proposes decentralized resource-aware coordination schemes for solving network optimization problems defined by objective functions which combine locally evaluable costs with network-wide coupling components. These methods are well suited for a group of supervised agents trying to solve an optimization problem under mild coordination requirements. Each agent has information on its local cost and coordinates with the network supervisor for information about the coupling term of the cost. The proposed approach is feedback-based and asynchronous by design, guarantees anytime feasibility, and ensures the asymptotic convergence of the network state to the desired optimizer. Numerical simulations on a power system example illustrate our results.