Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


Single-Agent Controller Design for Multi-Agent Environments

Many engineering systems can be viewed as an interconnection of interacting subsystems. The resulting overall model complexity can easily prohibit a centralized approach to controller design. A promising alternative is to design single-agent controllers with the forward-looking understanding that operations will occur in a multi-agent setting. In this way, controller design complexity primarily depends on the model complexity of a single agent, and, therefore, easily scales to large numbers of interacting components. This talk presents two complementary approaches towards this end. In the first approach, set-valued methods of constrained regulation are used to derive controllers that both assume and guarantee critical interaction levels, which are determined off line, that are mutually satisfactory among interacting agents. In the second approach, interacting agents self-organize in an online manner to derive controllers that are optimal from any single agent’s perspective. The methods are illustrated on various multi-agent examples, including vehicle platoons, manufacturing transfer lines, distributed resource allocation, and traffic routing.

Type of Seminar:
Public Seminar
Jeff Shamma
University of California Los Angeles, Los Angeles
Jan 30, 2006   16:30

ETH-Zentrum, Main Building, Room HG F 33.1, Rämistrasse 101, Zurich
Contact Person:

Prof. Lino Guzzella
No downloadable files available.
Biographical Sketch: