Learning and Evolution in Games: Adaptive Heuristics
H. Peyton Young · The New Palgrave Dictionary of Economics · 2008
‘Adaptive heuristics’ are simple behavioural rules that are directed towards payoff improvement but may be less than fully rational. The number and variety of such rules are virtually unlimited; here we survey several prominent examples drawn from psychology, computer science, statistics and game theory. Of particular interest are the informational inputs required by different learning rules and the forms of equilibrium to which they lead. We shall begin by considering very primitive heuristics, such as reinforcement learning, and work our way up to more complex forms, such as hypothesis testing, which still, however, fall well short of perfectly rational learning.