Semi-quantitative models in hybrid control systems

A typical hybrid control system is composed of a continuous state plant and a discrete-event supervisory system that assesses the state of the plant and generates and evaluates alternative policies for performing tasks, which are either scheduled operations or reactions to unforeseen events, such as failures. Modeling of such hybrid (mixed continuous and discrete-event) systems has received considerable attention and many models have been proposed already. We present here one alternative points of view that caters to the role of a supervisory control system as considered above. The base for the new model/paradigm is the use of contingent derivatives that permit a semi-quantitative prediction of future system behavior