SocSIM-K Social Simulation and Knowledge

SocSIM-K focuses on understanding human behaviours in complex real-world settings and building computational models of these behaviours, and their reasons behind them for simulation. Typically, we use an agent-based approach, grounding our work in the general area of agent-based social simulation. 

Agent-based social simulation is a scientific field dedicated to studying social phenomena through computer-based multi-agent models. In such models, individuals or groups are represented as intelligent autonomous agents whose interactions generate collective social dynamics. We draw heavily on social science theories, as well as empirical data, in order to conceptualise behaviours and people’s interactions with each other and with their environment. SocSIM-K focuses on the design and implementation of agents and their decision-making processes. By placing these artificial agents in simulated societies, we can see how individual actions lead to emergent social patterns. These simulations make it possible to explore scenarios that are difficult, impractical, too costly, dangerous, or unethical to study in the real world.
 

Research Challenges:

• How can we model factors which strongly affect our behaviours, e.g. social ties, cognitive biases, risk perception, emotions, etc. into agents?
• How can we develop cognitive computational models of human behaviour that represent real human behaviours?
• How can we validate our models/simulators?
• How can we collect/analyse data that are used to build our models?
• How can we deal with high computational expense?
• How can we ensure consistency in multi-level simulation (micro-meso-macro)?

Application areas: 

Crisis and emergency management, Cultural heritage, Pedestrian mobility in cities, Effect of environmental factors (pollution, energy, land management, climate) on cities, migrant trajectories.