Efficient Analysis of Discrete Event Systems: Supervisor Synthesis with Binary Decision Diagrams

Doctoral thesis, 2004

Due to the ever-increasing complexity of software and hardware, it is becoming more and more important to develop methods and tools for automatic analysis and synthesis of systems. One obstacle in achieving this goal has been the unmanageable size and complexity of the involved mathematical operations. For example, real-world systems such as manufacturing cells or embedded systems represented as composite discrete-event systems tend to have huge numbers of states. The enormity of these models will make most types of analysis very hard if not impossible. Taking advantage of Binary Decision Diagrams, this work presents a set of algorithms and heuristics for searching the enormous state-space of large discrete event systems in a efficient manner. Furthermore, the author demonstrates how important problems within the field of supervisor control theory can be solved using the presented search methods. It is also shown how the conjunction of such methods and algorithmic optimization is used to gain performance even more.