Contemporary process control includes continuous and discrete componen
ts. At the lowest level, continuous PID controllers are used for actua
tion and control. At a higher level, supervisory control mechanisms ar
e used to select appropriate control algorithms for the different mode
s of system operation to achieve optimal or near-optimal control. Mode
ling and analysis of such combined discrete and continuous components
requires hybrid modeling techniques. This paper presents a hybrid mode
ling paradigm, and discusses its execution semantics, which are based
on the principles of invariance of state and temporal evolution of sla
te. The modeling and simulation methodology is used to analyze the con
trol behavior of dynamic physical systems, and a model-verification te
chnique based on divergence of time demonstrates possible applications
in design tasks. (C) 1998 Elsevier Science Ltd. All rights reserved.