Using Game Theory for Distributed Control Engineering
Anders Rantzer · Lund University Publications (Lund University) · 2008
The purpose of this paper is to show how ideas from game theory and economics may play an important role for decentralized controller design in complex engineering systems. The focus is on coordination through prices and iterative price adjustments using a gradient method. We give a quantitative bound on adjustment rates that is sufficient to guarantee global convergence towards a Nash equilibrium of control strategies. The equilibrium is an optimal solution to the corresponding team decision problem, where a distributed set of controllers cooperate to optimize a common objective. The method is illustrated on control of a vehicle formation (e.g. automobiles on the road) where the objective is to maintain desired vehicle distances in presence of disturbances.