Managing complexity in large corporate information systems

Abstract

A control process is proposed based on managing a large number of influence parameters; using clustering/visualization concepts to manage the complexity associated with such large amounts of information; and considering as a reference the global knowledge of a distributed human control network. It is emphasized that using the strong capabilities of human visual reasoning in the analysis of complex systems may be an effective way of dealing with their complexity. An effective visualization process based on geometric 2-D/3-D representations of multidimensional spaces as well as clustering procedures constitutes the key to the process. Defining an appropriate reference model is also important and possible if a deep knowledge of the system is available. A simple demonstration model developed to evaluate some ideas of the proposed system is described.<>