Supervisor Localization
Kai Cai · Wiley Encyclopedia of Electrical and Electronics Engineering · 2019
Abstract In this article, we formulate a distributed control problem for multiagent discrete‐event systems, and introduce asupervisor localizationapproach to solving this problem. The approach defines acontrol coveron the state set of the supervisor, and constructs local controllers by merging the states residing in the same cells of the cover. The algorithm for supervisor localization is of polynomial complexity in the state size of the supervisor, and architectural and symbolic methods are explored to improve the algorithmic efficiency to tackle large‐scale multiagent systems. Further variations of the central concept of control cover are introduced, which expand the features of local controllers including time, partial observation, communication delay, and infinite behavior.