Hybrid Data-Driven Algorithms for Networked Multi-Agent Systems - Stability and Robustness

Abstract

The main goal of the proposed research is to develop analytical and constructive tools for the development of model-free and data-driven stochastic hybrid dynamical systems in the context of multi-agent systems, with stability, convergence, and robustness guarantees. To achieve this objective, we will rely on and enhance the PIs recent work on a general framework for stochastic hybrid dynamical systems and model-free hybrid control systems, which allows for non-unique solutions and thus the interplay between randomness and independent decision making in multi-agent systems. Our goal is to generate novel analytical tools for the analysis of data-driven algorithms in the context of stochastic hybrid dynamical systems, as well as constructive procedures for the design of robust model-free and data-driven algorithms, with performance guarantees, for networked multi-agent dynamical systems.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 18, 2022
Accession Number
AD1230398

Entities

People

  • Andrew R. Teel

Organizations

  • University of California, Santa Barbara

Tags

Fields of Study

  • Computer science

Readers

  • Robotics and Automation.
  • Statistical inference.
  • Systems Analysis and Design