Alex Voss, School of Computer Science, University of St Andrews, UK
Rob Procter, Manchester e-Research Centre, University of Manchester, UK
Andy Turner, School of Geography, University of Leeds, UK

Agent-Based Social Simulation is a relatively new social science research method. It allows social phenomena to be investigated by studying dynamic processes in computational models of populations. Social simulation complements existing social science research methods and builds on empirical findings and theories of social phenomena. Applications of social simulation range from the study of basic demographic processes to uses in a range of interdisciplinary research areas such as public health, urban studies or criminology, to name a few examples.

Social scientists as well as other professionals such as policy makers have used various methods in the past to model social phenomena and generate predictions about them. At its simplest, one can extrapolate from empirical data by fitting functions that approximate the data gathered. In the past, a key limitation for such extrapolations was that the underlying model was a static one, defined in terms of mathematical formulas that could be manipulated manually. The necessary simplifying assumptions often meant that the applicability and the predictive power of these models was limited. The advent of computers and their increasing capacity has made it possible to build models that consist of representations of social agents and their environment and to simulate their behaviour and interactions.

Such agent based models (ABMs) are still limited by the capacity and processing speed of today's computer systems. It is now possible to simulate relatively large populations with a reasonable number of variables even on a desktop computer. However, to study social phenomena it is often necessary to run multiple simulations that vary specific variables in order to study their impact. Such parameter studies can be relatively easily distributed across a number of computers using grid technologies. Each instance of a simulation run with a specific set of parameters can be sent to a different machine in a grid, with all the machines working in parallel to produce an overall result.

AIMS:
This tutorial aims to provide an introduction to the principles behind social simulation and to demonstrate these principles by teaching participants how to develop a social simulation model and running a simulation using the Repast Simphony Toolkit. Building and running this model will involve the creation of a model population on the basis of freely available social science datasets. In order to investigate hypotheses or to test the impact of policy interventions of real-world events, it is often necessary to run ensembles of simulation jobs. Running such ensembles is most effectively done on high throughput computing infrastructures such as grids or clouds. Participants will learn about grids and clouds and will start using them by submitting ensembles of the models on real-world infrastructures.

Prerequisites:

• Participants should bring their own laptops to the tutorial if possible. This will allow them to continue using the software used in the tutorial and to continue experimenting with their models. A limited number of machines can be made available for those who cannot bring their own machine and notify the organisers in advance.
• No familiarity with social simulation tools or grid computing is required. Programming skills are also not required as all code for the simulation models used will be provided.
• Participants will be given access to all necessary software and data as well as a grid certificate required to access the grid infrastructure.

Further Details
Further details on the tutorial, the agenda for the day, instructions and links to the necessary software will be made available at:
https://e-research.cs.st-andrews.ac.uk/SimISGC2010

Time Media Conference Room 09:00 ~ 09:30 Registration 09:30 ~ 10:00 Welcome, What is social simulation? What is Agent-Based Modeling (ABM)? 10:00 ~ 10:30 The RePast Simphony Toolkit, an example model 10:30 ~ 11:00 Coffee break 11:00 ~ 12:00 Practical I - installing RePast and running model 12:00 ~ 12:30 Population reconstruction (Andy Turner) 12:30 ~ 14:00 Lunch 14:00 ~ 14:30 Infrastructures for Social Simulation (Rob Procter) 14:30 ~ 15:00 Introduction to grids and cloud computing 15:00 ~ 15:30 Coffee break 15:30 ~ 16:45 Practical II - running model ensembles on grids and clouds 16:40 ~ 17:00 Closing remarks 17:00 End

• The time, content and speakers might change.