Game theoretic approach to reconfiguration decisions by heterarchical machine controllers

Heterarchical control systems have emerged as promising alternatives to conventional centralized and hierarchical shop floor control structures. However research to date has largely ignored long term control issues such as the timing of resource reconfigurations issues vital for the avoidance of myopic decision making. As traditional analysis techniques implicitly assume centralized analysis and implementation of policies, they are not directly applicable to heterarchical environments. This paper describes a non-cooperative game theoretic planning technique for the autonomous analysis of reconfiguration decisions by heterarchical machine controllers. Under the proposed system, a machine controller evaluates a setup game to determine appropriate actions when faced with a reconfiguration decision. The prescribed action is defined by a pure or mixed strategy equilibrium of the setup game. By evaluating a game, the machine controller accounts for the potential actions and reactions of other machine controllers in the system. The approach allows autonomous machine controller to make strategic decisions regarding future actions while preserving the independence characteristic of controllers in heterarchical environments.