We discuss the theoretical foundations of distributed large-scale control. Our approach is to integrate multiagent microeconomic market theory with control theory, such that online adaptive control is economically optimal at the overall systems level. A central result is the derivation of a general market theorem that proves two important properties about this agent-based microeconomic control:(1) computational economies with dynamic pricing mechanisms are able to handle scarce resources for control adaptively in ways that are optimal locally as well as globally (‘societally’);(2) in the absence of resource constraints the total system acts as collection of local independent controllers that behave in accordance with conventional control engineering theory. We further derive some analytical results for the market-based control outcome, the clearing price, and hierarchical versus decentralized emergent control.
|Title of host publication
|Proceedings of CRIS 2004-2nd International Conference on Critical Infrastructures
|Published - 2004
|Securing Critical Infrastructures CRIS - Grenoble, France
Duration: 25 Oct 2004 → 27 Oct 2004
|Securing Critical Infrastructures CRIS
|25/10/04 → 27/10/04