Development of a Comprehensive Theory for Discrete Event Systems

离散事件系统（DES）用于模拟人工系统，这些系统不容易用传统控制理论中使用的微分方程或差分方程来描述，而是通过记录系统中重大质的变化的事件的相互作用来描述。已经提出了几种对DES进行建模、控制和优化的方法，每种方法都解决了一个特定的问题，并提供了DES整个谜题的重要部分。然而，这些部分在很大程度上是分开和断开的。此外，一些方法的计算复杂度呈指数级增长。为了克服这些障碍，本提案的目的是发展一个全面的DES理论。在这个理论中，将提供在线控制的有效实施；组件故障诊断；严格的时间限制将得到满足；时间性能将被评估；混合动力系统也会被考虑。如果成功，这一综合理论将弥合DES中两种主要方法之间的差距：解决逻辑问题的定性方法和解决性能问题的定量方法。它还将把离散事件系统的研究成果与连续变量系统的研究成果联系起来。它将进一步提供有效的计算方法来实施控制。所有这些都将大大增强DES方法的适用性。＊＊＊

9315344 Lin Discrete event systems (DES) are used to model man-made systems that are not readily described by the differential or difference equations used in traditional control theory, but rather by interactions of events that record significant qualitative changes in the systems. Several approaches to modeling, control, and optimization of DES have been proposed, each addressing a specific issue and providing an important piece of the entire puzzle of DES. However, these pieces are as yet separate and disconnected to a large extent. Furthermore, some approaches suffer from exponential computational complexities. To overcome these obstacles, the aim of this proposal is to develop a comprehensive theory of DES. In this theory, efficient implementation of on-line control will be provided; component failures will be diagnosed; hard time constraints will be satisfied; temporal performance will e evaluated; and hybrid systems will be considered. If successful, this comprehensive theory will bridge the gap between two major approaches in DES: qualitative approach addressing the logic issues and quantitative approach addressing performance issues. It will also connect research results in discrete event systems with those in continuous variable systems. It will further provide computationally efficient ways to implement control. All these will greatly enhance the applicability of DES methodologies. ***