Scalability and Implementation Issues for Discrete-Event Systems

点击翻译按钮获取中文摘要

The primary objective of this proposal is to address the two main concerns standing in the way of the adoption of DES theory by industry: scalability and implementation issues. In discrete-event systems (DES), a great deal of effort is currently being applied to overcome the state space combinatorial explosion problem, in order to allow supervisory control methods to scale to large systems (> 10^20 states). To deal with the complexity of large scale systems, the software engineering community has long advocated the decomposition of software into modules (components) that interact via well defined interfaces. The applicant has developed a similar approach, called hierarchical interface-based hierarchical supervisory control (HISC). This approach develops well defined interfaces between components to provide the structure to allow local checks to guarantee global properties such as controllability or nonblocking. The proposed research will build upon this framework by extending and generalising it, adapting existing DES methodologies so that they can exploit the HISC architecture and developing tools that utilise the method. In particular, the research will adapt the HISC method to timed DES, DES fault-diagnosis, and decentralized supervisory control. Also, the HISC concept will be extended to other software verification methods, such as state-based specifications verified using theorem provers. DES also needs a reliable method to transform a DES supervisor into an actual software or hardware implementation. In doing so, concurrency and timing delay issues must be resolved. This proposal will build upon the applicant's Sampled-Data (SD) Supervisory control work. First, a method to automatically generate software code or hardware design from SD supervisors will be developed. Next, SD control will be extended to the HISC platform. Then, a discrete-optimization method will be developed that respects the SD conditions. Finally, this method will be adapted to the HISC platform. The above tasks are key to making DES ready for industry adoption.