Uncertain Behavior

The invited session on behaviors, modeling, and robust control reeects an emerging view that these apparently disjoint subjects have important connections that can have a major impact on both theoretical and practical aspects of control. The behavioral setting provides a convenient framework for connecting modeling and robust control, and is being extended to make deeper contacts with more mainstream control. 1 Why Behave? In an attempt to capture the essence of the behavioral paradigm, one would do quite well with the following: the absence of a preconceived notion that dynamical systems must be described as input-output maps. Invariably, the majority of texts that deal with dynamical systems take the de facto stance that systems are inherently input-output maps. A brief excursion to the bookshelf yielded the following: From Signals and Systems, 13]: A system can be viewed as any process that results in the transformation of signals. Thus, a system has an input signal and an output signal which is related to the input through the system transformation. A system is a mathematical model or abstraction of a physical process that relates inputs or external forces to the output or response of the system. It is not necessary in the behavioral setting to consider systems as operators that transform inputs into outputs. Rather, all variables are considered a priori on an equal footing, and the input-output framework is developed as a special case which in many situations can actually be deduced from the original model. It may seem that the above is simply a generalization for its own sake, and adopted solely to allow for greater levels of abstraction. This is a reasonable interpretation if we view 1