Please note: this event has passed
Abstract
We consider systems in which dynamic behaviour is spatially distributed and also stochastic. In particular, the agents within the system have a location, and the ability of agents to interact may depend on agents being co-located or within a given range of each other. Such systems are typically studied by simulations. Current simulation-based approaches provide summary information about the satisfaction of properties over the spatial domain, providing estimated values that include intrinsic uncertainty.
In this work we seek to take this uncertainty into account and to enrich the summary information through the use of a novel logic, the Three-Valued Spatio-Temporal Logic (TSTL), which allows us to reason, not only about the behaviour of the system, but also about the evolution of the satisfaction of properties expressed in a spatio-temporal logic. This provides additional insight into the dynamic behaviour of the system under study. For example, in the analysis of the efficacy of a control measure for fire spread, we can verify whether the spread in a specific area will happen with probability under a given threshold over time. We can also identify the locations that are at highest risk, because they are surrounded by locations with high probability of burning. We use a three-valued logic, with the third value unknown indicating when there is insufficient evidence to make a judgement about a property. This can be taken as an indication of when more simulation trajectories are needed to evaluate propositions more precisely.
Joint research with Ludovica Luisa Vissat, Michele Loreti, Laura Nenzi and Glenn Marion.
Biography
Professor Jane Hillston is a Professor of Quantitative Modelling and Head of School in the School of Informatics within the College of Science and Engineering, and Deputy Vice Principal for Research at the University of Edinburgh. Jane is a Fellow of the Royal Society of Edinburgh and a Member of Academia Europaea. She currently chairs the Executive Committee of the UKCRC (UK Computing Research Committee). She is a member of the Informatics Europe working group on Women in Informatics Research and Education, and is a member of the LFCS (Laboratory for Foundations of Computer Science) in the School of Informatics. In 2004, Jane was the recipient of the first Roger Needham award. In 2018, she received a Suffrage Science Award. Jane’s research is concerned with modelling and reasoning about the dynamic behaviour of systems, using quantified formal methods. Applications have ranged from cellular processes to transport systems.
How to join
The seminar can be joined through Microsoft Teams - either through the Teams app or your web browser.