About The Forum
Calendar 2009 Poster Sessions : Reduction of Decentralized Control Problems to Tractable Representations
Abstract:
We address the question of which decentralized control problems are tractable. The largest known class is the quadratically invariant systems. We broaden this set by characterizing a class of systems which are not quadratically invariant, but reduce to systems that are. We call this "internal quadratic invariance". We give an algorithm for performing this reduction, and illustrate our method with examples.
Bio:
Laurent Lessard was born in Toronto, Canada. He is now working with Sanjay Lall in the Networked Systems lab at Stanford. His research interests include: linear systems, multiscale representations, and decentralized estimation and control.
We address the question of which decentralized control problems are tractable. The largest known class is the quadratically invariant systems. We broaden this set by characterizing a class of systems which are not quadratically invariant, but reduce to systems that are. We call this "internal quadratic invariance". We give an algorithm for performing this reduction, and illustrate our method with examples.
Bio:
Laurent Lessard was born in Toronto, Canada. He is now working with Sanjay Lall in the Networked Systems lab at Stanford. His research interests include: linear systems, multiscale representations, and decentralized estimation and control.