Distributed simulation and control: the foundations

This paper seeks a new simulation and execution paradigm for the design and operation of complex systems. An expanded life cycle for a simulation model is first provided. It is assumed that complex systems can be represented as systems of interacting subsystems, which evolve by executing tasks upon objects. Care is taken to distinguish the real world where process execution occurs from the virtual world where planning is addressed. It is illustrated that the ideal model should be able to both evaluate and control the subsystem that it addresses. The advantages of such approach are discussed with relation to both validation and execution needs. In particular, it is demonstrated that a distributed-controller based paradigm could provide significant advantages in the evaluation of the system using distributed simulation. This form of execution is also contrasted to evolving on-line simulation requirements that will support the realtime distributed management of these systems