Process Windows

摘要

We describe a method for formally representing the behaviour of complex processes by process windows. Each window covers a part of the system behaviour, i.e. a part of the underlying transition system, and is easier to understand and analyse than the complete transition system. Process windows can overlap and have shared states and transitions so that the complete system behaviour is the union of window behaviours. We demonstrate the advantage of such representations when dealing with complex system behaviours, and discuss potential applications in circuit design and process mining. As a motivational example we consider the problem of covering transition systems by marked graphs, or more generally choicefree Petri nets. The obtained windows correspond to choice-free behavioural scenarios of the system, wherein one window can take over, or wake up, after another window has become inactive. The corresponding wake-up conditions and wake-up markings can be derived automatically.