DEDS控制优化及其在FMS中的应用研究

Discrete event systems are a class of event-driven dynamic systems. There is an important application of discrete event systems in the flexible manufacturing systems, computer communications systems, and traffic systems. Its control problem is one of the main problems researched in this area. The project models the discrete event systems by Petri nets, analyzes the liveness of the systems and control problems, and establishes the system controllers with some special properties. We presented the concept of the path gain in Petri nets, and a method for synthesizing the Petri net invariants by analyzing path gains. Then by implementing these invariants we can obtain a Petri net controller for the systems with uncontrolled transitions. The controller guarantees that the closed system is live. Here importance is that the mathematic programming and invariant are combined into together, a minimum limited Petri net controller is synthesized and a new method for synthesizing controller is presented..Address the deadlock avoidance problems in manufacturing system with sequence and route flexibility. It is shown that for a large class of such systems, through one-step look ahead, an optimal deadlock avoidance policy can be obtained. For a general system we present a system reduction method by which the system is reduced and the reduced system has an optimal deadlock avoidance policy. Using this optimal control policy and restricting the allocation of resource with capacity one, a sub-optimal deadlock avoidance for the system is obtained..A hybrid system is a combined system of discrete event system and continual systems. In the project the stability of sampled data control systems, typical hybrid systems, is studied. A Lyapunov uniformly asymptotically stable condition and an unstable condition are characterized. It is proved that if the matrix of the linearized system of a non-linear system is schur stable, then the system is Lyapunov uniformly asymptotically stable. Utilizing the concept of cycle invariant subspace, the controllability and observability of a class of linear switching system are studied. Necessary and sufficient condition on the controllability and observability of the system are obtained respectivey.In the project, a double iterative algorithm of optimal control for continuous-time nonlinear system is proposed on basis of a bilinear model, the convergence of the algorithm is also proved. For discrete-time nonlinear time-delay system, a dynamic integrated system optimization and parameter estimation method is constructed, a sufficient condition on the convergence of the method is given and the optimization of the converged solution is also proved. An adaptive RBF neural network control approach is presented for a class of nonlinear system with unknown nonlinearities. Semi-global ultimate uniform boundedness of the resulting adaptive system is proved. For parametric time-delay output-feedback nonlinear system, an adaptive control strategy is proposed by recursively constructing a Lyapunov-Krasovskii functional. A sufficient condition on ensuring asymptotic tracking property of the closed-loop system is also obtained. In view of adaptive control, two adaptive iterative learning control strategies are designed from time domain for both strict-feedback system with unknown nonlinearities and parametric output-feedback nonlinear time-delay system. Utilizing the stability theory of hybrid system proves the stability and convergence of the overall systems. Moreover, exactly tracking of objective trajectories are also obtained.

离散事件动态系统的控制优化是自动化领域的重要研究课题之一。本项目研究离散事件动态系统的逻辑、实时控制策略及其综合算法，建立实时系统的性能优化与评估方法，研究实时离散事件系统满足控制问题要求的各种调度策略及其综合方法。所提出的理论与方法将有助于推动离散事件系统理论及其应用的发展，在计算机集成制造特别是柔性制造系统，混杂动态系统，计算机网络，交通管理等系统中都有重要的应用价值。