Object Oriented Design of a Multilevel Hierarchical Organization Structure

Multilevel hierarchical control has been placed in a well-defined conceptual and mathematical framework in the sixties and culminated in the publication of the book by Mesarovic et al. (1970). The concept of hierarchy was clarified by identifying three notions of levels: • The organizational level, the echelon. Much of the work in control systems and in organization theory deals with the first two types or with a combination of both of them. For example, the hierarchical decision structure of Valavanis and Saridis (1992) can be thought of as a layered one, which is mapped onto a hierarchical physical architecture of the echelon type. The layers support the concept of Increased Precision with Decreasing Intelligence. (Saridis, 1989). It is a particularly effective conceptualization, especially when considering systems where hardware and software are closely integrated such as intelligent robotic systems. A new problem that has began to appear recently-at least at the conceptual level-is the design of the information architecture in the context of which a number of intelligent entities operate.(Levis and Perdu, 1994) One can envision several autonomous robotic systems that are asked to operate in a coordinated manner, whether through centralized or decentralized coordination. The passing of information between the different entities and the need for a flexible, adaptive design of the coordination schemes lead to a software intensive system design. Consequently, a different way of looking at the total system is appropriate-the stratified view. The concept of a stratum is used for modeling information architectures when viewed from different levels of abstraction or, alternatively, of refinement. Zaidi (1991)has already used the approach to model and analyze a type of multilevel, hierarchical system defined as multiechelon stratified systems, where the system is described by a family of structures each concerned with the behavior of the system as viewed from a different level of refinement. The intelligent decision