Hierarchical simulation of timed behaviours of structured occurrence nets

Abstract

A complex evolving system (CE-system) is composed of a large number of (sub-)systems concurrently interacting among themselves and with the system’s environment. A CE-system is subject to modifications by other systems. Typical examples of CE-systems are large distributed systems whose software is continually updated, or dynamically evolving (cy- ber)crime investigations. Time simulation is a powerful tool used in many fields to model and analyse the behaviour of CE-systems, such as crimes and accidents. In crime investigations, time simulation may be an effective technique since it enables investigators to piece together the sequence of events that led to a crime. Investigators may better grasp the circumstances leading up to a crime by simulating crucial events happening in time, which can be useful in identifying suspects and acquiring evidence. A typical notation for recording the behaviour of a CE-system is some form of a directed acyclic graph. The framework of structured occurrence nets (SO-nets) can play an important role in the representation of CE-system behaviours. In general, SO-nets are sets of related acyclic Petri nets, employing different types of formally defined relationships and supporting various types of hierarchy and abstraction, which represent the details of concurrency and causality relations between executed events. This thesis focuses on two objectives. First, it seeks to develop theoretical underpinnings, new algorithms, and implemented prototype software tools for hierarchical and abstraction- based analyses and simulations of timed behaviours within CE-systems represented using SO-nets. Second, it aims to analyse scenarios in which different time granularities are utilized at different levels of abstraction. This involves developing theoretical frameworks, algorithms, and prototype software tools tailored to aforementioned specific requirements. The discussion is carried out using behavioural structured acyclic nets (BSA-nets) that are part of the SO-nets framework. The overall research question to be addressed is whether the resulting framework can provide effective support for incident investigators.