Odern Control Systems Must Meet In- Creasingly Demanding Requirements

creasingly demanding requirements stemming from the need to cope with significant degrees of uncertainty, as well as with more dynamic environments, and to provide greater flexibility. This, in turn, means that control systems software is highly complex in that it invariably has a large number of interacting parts [20]. This complexity requires that state-of-the-art software engineering methods and techniques be employed. In this article, we will argue that analyzing, designing, and implementing such complex software systems as a collection of interacting, autonomous, flexible components (i.e., as agents) affords software engineers several significant advantages over contemporary methods. In seeking to demonstrate the efficacy of the agent-oriented software engineering approach [11], the most compelling argument would be to show quantitatively how its adoption improved the development process in a range of (control system) projects. Although several applications have been deployed (see [15] and [19] for overviews), such data are simply not available (as is the case for other contemporary software engineering approaches such as patterns, application frameworks, and