Coordination of Distributed Control Systems

A general heuristic methodology to specify coordination algorithms for distributed systems is presented. The methodology is summarised in six fuzzy rules. To translate these rules into effective tractable algorithms, two methods will be examined: via a local coordination and, in terms of Max Weber: an 'ideal behaviour'. 'We demonstrate that thereby principally different types of systems are obtained. Often, coordination can only be achieved by the second method and the behaviour of those systems shows phenomena that are impossible in locally coordinated systems: downward causation and strong anticipation.