Deterministic dynamics of distributional multi-agent reinforcement learning

The useful part

by Clémence Bergerot, Pawel Romanczuk, Wolfram Barfuss Understanding how cognition shapes behavior across contexts remains a fundamental challenge for many disciplines. In particular, for the optimism heuristic–i.e., the tendency to overweight positive (relative to negative) information–knowledge remains fragmented, with models developed in specific domains in isolation. Here, we present a unifying computational framework by deriving the deterministic dynamics of distributional multi-agent reinforcement learning.

How it works

• We validate our framework by reproducing established results across three iconic domains spanning individual bandit choice under resource variability, social coordination, and risky choice.
• We further reveal "individual dilemmas": circumstances where agents gravitate toward suboptimal yet stable strategies, offering a mechanistic explanation for incoherent choice patterns.