Planning and Coordination in Hierarchies of Intelligent Dynamic Systems

Abstract

On the basis of the known principle of interactions prediction (Mesarovic), our earlier proposed incremental coordination principle is extended over hierarchical collectives of intelligent dynamic systems (IDSs) after Gennady Osipov. Such systems admit arbitrary types of variables in their state vector and thereby allow investigating more general dynamic systems than “classical” ones defined in numerical state spaces. Using the concept of effective N-attainability (Osipov), a straightforward procedure of planning for hierarchical collectives of IDS is developed. As soon as a plan for reaching a goal state from the current one is found, effective implementation of this plan requires for coordination of IDSs taking their parts in the collective. We consider both aspects of coordination (coordinability with respect to the coordinator’s task and coordinability in relation to the global task) and infer necessary conditions of the coordinability for a locally organized hierarchy of IDSs.