Learning from Gamettes: Imitating Human Behavior in Supply Chain Decisions

Gamettes are playful tools for agent-based participatory simulation and have shown to be valid for collecting rich behavioral data from human decision-makers. However, there is still a question that how such data can be used to create or update agent-based and behavioral models. In this paper, we evaluate and compare the performance of different methods for imitating human behavior. We use extracted data from gamettes in an empirical study on supply chain decisions, and compare the performance of a nonlinear regression model with two imitation learning algorithms. Our results demonstrate that each method is capable of modeling and thus predicting human behavior by considering multiple trajectories from different players.