Modeling and design of interacting continuous-time/discrete-event systems
Abstract
This research addresses problems associated with the modeling and design of interacting continuous-time system (CTS) and decision-making system (DMS) called hybrid system (HS). HSs appear each time a continuous-time process has to be supervised by a discrete-event decision-maker for attaining a number of goals. An example of such a scenario is the modeling of flexible manufacturing systems, which have motivated the introduction of HSs. A graphically expressive Petri Net (PN) subsumes the role of the DMS while a state model (SM) represents the CTS. Two event-based interfacing functions communicate information between the two components of the HS. The transition enabling function (TEF) maps events in the CTS, events in the DMS and numerical processor events into a pattern that controls the firing of transitions and thus the evolution of information flow in the PN representation of the DMS. The actuator function (AF) maps events in the DMS into a number of actuators altering the structure of the CTS. The action of the TEF on the PN corresponds to the creation of a family of PNs indexed by the range of the TEF, while the action of the AF on the SM representation of the CTS corresponds to the creation of a family of CTSs indexed by the range of the AF. The modeling methodology is developed through two examples, a fishery resource management example and a continuous-time m-switched system example. In the latter case the state-space is partitioned into a number of energy-reducing regions, one associated with each system, whose union covers the state-space. The existence of suitable switching which produces a discontinuous energy reduction on the system guaranteeing its global asymptotic stability. The switching strategy is decided by the supervisory discrete-event system. Moreover, the projection of discrete-time switched systems into PNs allows for the analysis of the symbolic dynamic properties of the switched system which are further exploited by a supervisory decision-making system.
Degree
Ph.D.
Advisors
DeCarlo, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.