Cognitive Modelling

In my bachelor’s thesis, I modelled social cognition in a public-goods game to simulate diverse cooperative group behaviour and enable data-sparse yet psychologically grounded forecasts.

I built a multi-agent simulation with pyactr, where heterogeneous agents chose their contributions via a social-norm utility function. This function also grounded mental models of reputation, consistency and consensus.

I varied multipliers, population sizes and punishment mechanisms, generated synthetic datasets and compared the results with empirical lab studies to validate representational accuracy and robustness.

The results show how psychologically anchored micro-processes in agent models can produce realistic emergent group patterns at the macro level.

Technologies: Python, ACT-R, pyactr, Multi-Agent Simulation