Skip to main content

Learning Decision-Making Models Using Non-traditional Equilibrium Concepts

Game theoretic models often assume complete rationality. However, in practice decisions are often made myopically (bounded rationality). We are developing new equilibrium models of decision-making in a modified game theoretic context using decision-making processes with reduced notions of rationality. With these equilibrium concepts, our research builds on our existing work in the area of utility learning and inverse optimization to create a relaxed inverse game theory framework for learning decision-making models from observed behavior.