Intelligent modeling for control, reconfiguration and optimization of discrete event systems

The main objective of this work is to develop an integrated design methodology for modeling, control, reconfiguration, and optimization of discrete event systems. The application focus is on automated manufacturing systems which are characterized by both discrete and continuous variable dynamics. Fuzzy sets are introduced in representing the states of the system. The processing times and setup times of different machines and workstations in a manufacturing system are not deterministic in many cases. These are represented by fuzzy distributions. The use of fuzzy logic assists in representing the status of a plant more realistically and results in prompt controller decision with less time delay. An extended fuzzy Petri net is defined and its properties related to manufacturing systems are presented. An equivalence of the modeling approach with the extended fuzzy Petri net is established. The generality of the modeling and control approach is established by applying it to different popular classes of manufacturing systems which include but are not limited to flow shops and job shops with re-entrant lines. This indicates the power of the approach as compared to some of the existing approaches which are severely limited to some special classes of manufacturing systems. A system reconfiguration scheme is developed which operates in case of some component failure or in case of operational changes. The operational changes may include a change in objective criteria or a change in production plan. The reconfiguration sche systems which include but are not limited to flow shops and job shops with re-entrant lines. This indicates the power of the approach as compared to some of the existing approaches which are severely limited to some special classes of manufacturing systems. A system reconfiguration scheme is developed which operates in case of some component failure or in case of operational changes. The operational changes may include a change in objective criteria or a change in production plan. The reconfigensive job shop problem is presented where the reconfiguration scheme clearly demonstrates its effectiveness in terms of selection of routes that best satisfy the objective functions.;The major contribution of the thesis are: (1) Development of an integrated design methodology that addresses modeling, control, reconfiguration and optimization in a unified manner; (2) Presentation of a hierarchical modeling procedure that captures hybrid discrete/continuous dynamics of the system; (3) Introduction of fuzzy sets in representing states of the system; (4) Presentation of a system-theoretic approach to control, where controller acts as a discrete event scheduler; (5) Defining a generalized fuzzy Petri net and establishing its equivalence with the modeling approach; (6) Presentation of different application examples from a variety of manufacturing configurations--establishing generality of the modeling and control approach; (7) Development of an architecture for system reconfiguration that is based upon fuzzy comparison of competing objectives. (Abstract shortened by UMI.).