The PI does research on the modeling, analysis, and control of discrete event systems (DES's). A DES is a man-made dynamical system whose behavior is described by the complex interactions of distributed, communicating, controlled processes. These interactions are modeled by traces of events that record significant changes in the state of the system. A process may for example be a computer program or a user's transaction in a computer system or a machine or a robot in a flexible manufacturing system. The first component of the research deals with theoretical work on the control of DES's. Specific issues being studied include: detailed study of the properties of controllers for DES's; the synthesis of controllers with good qualitative and quantitative performance characteristics; modeling and analysis of recovery in the control of systems that block; and on-line supervisory control under partial information. Progress on these problems will enhance the scope and applicability of the emerging control theory for DES's. The second component of the research deals with the development of algorithms for controller synthesis problems in a relational algebraic framework. This approach is pursued in order to address synthesis problems for large scale DES's. Finally, representative case studies from the fields of database systems, manufacturing systems, and traffic systems are being analyzed in conjunction with the above development of theory and algorithms.
