Agent-based distributed dependable systems

对基于软件的系统日益重要的需求是能够及时地交付可靠、正确和安全的服务。这种能力通常被称为系统可靠性，它通过自我管理、自我维护和自我修复/自我修复等特性来证明。这些特性正在成为许多工业部门对基于软件的系统的要求，例如，在一些制造环境（如钢铁厂和化工厂）和关键的基础设施/设施（如电网和核电站）中，系统故障或意外关闭将导致重大的经济损失，甚至破坏整个工厂/设施。本研究的目的是发展科学和工程基础、创新技术和实用的解决方案，以支持分布式、自动化、可靠的计算机系统，这些系统能够实时处理故障和不确定的变化，而不会（或减少）服务的退化。重点是在动态制造环境和关键基础设施/设施中理解、建模、设计和实施复杂的软件系统。我们将特别研究智能软件代理在这一领域的应用，通过与其他概念、原理和技术相结合，包括基于活动事件的概念、最优控制理论、调度和优化、数据挖掘和分布式传感器网络。

An increasingly important requirement for software based systems is the ability to deliver reliable, correct, and safe services in a timely manner. This kind of ability is usually called system dependability which is demonstrated through features like self-managing, self-maintaining, and self-repairing/self-healing. These features are becoming a requirement of software based systems in many industrial sectors, e.g., in some manufacturing environments (such as steel mills and chemical plants) and critical infrastructures/facilities (such as power grids and nuclear power plants) where system failures or unexpected shutdowns will cause significant economic losses or even destroy entire plants/facilities. The objective of this research is to develop scientific and engineering foundations, innovative technologies, and practical solutions to support distributed, automated, dependable computer systems that are able to handle faults and uncertain changes at real time without (or with reduced) degrading of services. The emphasis will be on understanding, modeling, designing, and implementing complex software systems in dynamic manufacturing environments and critical infrastructures/facilities. We will particularly investigate the application of intelligent software agents in this area, by integrating with other concepts, principles, and technologies, including active event-based concepts, optimal control theory, scheduling and optimization, data mining, and distributed sensor networks.