Triggered Control for Distributed Optimization and Learning in Networked Multi-Agent Systems
Abstract
This proposal has laid down the foundation of a novel framework for the design of distributed triggered control strategies that endow networked systems with greater autonomy and decision making capabilities in dynamic environments subject to uncertainty and evolving task specifications. Our approach has combined the reactive nature of event-triggered control with the autonomous features of self-triggered control in providing algorithmic solutions to the scenarios of distributed optimization in cooperative networks and distributed learning under networked strategic interactions. The conceptual novelties of the project hinge upon the notion of agent abstractions and promises about future states, an original combination of event- and self-triggered information updates, new methods to reason and operate on set-valued information models, and new techniques for distributed controller design and analysis.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 07, 2020
- Accession Number
- AD1107049
Entities
People
- Jorge Cortés
Organizations
- University of California, San Diego