The session

The goal of this session at SocSimFesT is to share the lessons learnt by all researchers that have been involved in modelling the COVID-pandemic one way or another, to make our social simulation community more prepared for such events in the future. Challenges that will be addressed include: How to build fast development of empirically grounded simulation model of human behaviour, How to build general models of human decision-making and behaviour that can easily be adapted and integrated (e.g. Josh Epsteins approach on fast crisis response using prepared relatively specific). How do we determine what is important and what is not? E.g. in ASSOCC we do not have a real spatial model. It is a limitation for some questions. Is that ok? Or does it make the platform not usable in most crises? How can we explain the results of a simulation promptly and clearly to stake holders? In a crisis we have to justify results. Especially if they are unexpected. How can this be supported in an efficient way?

About ASSOCC:
During the COVID-19 pandemic a group of self-driven European researchers came together to build an agent-based model that could contribute to the control of this devasting world crisis. The model, which was named ASSOCC was a great success in providing societal and policy insights, but perhaps a more important outcome was the learning process and the hurdles that were not known beforehand. For example, during the development of the model, new situations and dimensions emerged that were completely hidden at first. This called for changes, additions, deletions of factors that were dependent on the developments and experiences in the real world. As another example, in crisis situations a quick and on demand response is essential which is not an experience modelers are acquainted with. These both suggest the need for flexible modelling platforms that can change on demand.

Program

Throughout the workshop, we will be using Mural (https://www.mural.co/) to collect discussion notes. Per session, there will be a separate mural link. We hope to draft an opinion paper as an outcome of this workshop

Session 1: Simulation models during crisis (Lois Vanhee)

9:40 – 9:50 Presentation (Jason Thompson)

9:50 – 10:15 Plenary session or breakout rooms depending on the number of audience (see discussion points below)

Session 2: Abstract models vs data-driven complex models (Frank Dignum)

10:15 – 10:30 Presentation (Frank Dignum)

10:30 – 11:00 Plenary session or breakout rooms depending on the number of audience (see discussion points below)

11:00 – 11:15 Coffee break

Session 3: Social simulation frameworks for crisis situations (Amineh Ghorbani)

11:20 – 11:30 What agent-based modelling needs to have prepared by the next pandemic (Christopher Watts)

11:30 – 12:00 Plenary session or breakout rooms depending on the number of audience (see discussion points below)

Session 4: Model validation, comparison with other models, and comparison with real-world (Fabian Lorig)

12:00 – 12:20 presentation (Fabian Lorig)

12:20 – 12:45 Plenary session and Wrap up

Organisers

Amineh Ghorbani, TU Delft (contact person: a.ghorbani@tudelft.nl)

Frank Dignum, Umeå University

Harko Verhagen, Stockholm University

Fabian Lorig,  Malmö University

Paul Davidsson, Malmö University

Loïs Vanhée, Umeå University