Stefano Blando.
Italiano

01 — Research

Agents

Who acts: decision-makers that learn, adapt and decide, from agent-based models and multi-agent reinforcement learning to LLM-based agents and the algorithms now active in markets.

Economic outcomes emerge from the choices of many heterogeneous decision-makers: households, firms, banks, regulators and, increasingly, algorithms. This pillar asks which behavior a model should assume and how that behavior changes when the environment does. It combines agent-based models with bounded rationality and heuristic rules, multi-agent reinforcement learning, and agents that reason with large language models, to study how local adaptation shapes aggregate dynamics.

Projects