Complexity, Evolution and Learning: a simple story of heterogeneous expectations and some empirical and experimental validation

This note discusses complexity models in economics. A key feature of these models is that agents have heterogeneous expectations, disciplined by adaptive learning and evolutionary selection. Agents adapt their rules based upon past observations and switch between different forecasting heuristics based upon strategy performance. We discuss how these models match empirical facts as well as laboratory experiments with human subjects and how this approach may tame the ``wilderness of bounded rationality''