Multi-agent LLM simulations testing algorithmic collusion and coordination breakdown in oligopoly markets.
Investigating whether LLM algorithmic price collusion survives agent replacement through inherited price history and tacit coordination.
Learning agents in oligopolies (Cournot / Stackelberg) Agent-based model
Replication of Calvano et al. (2020) 'Artificial Intelligence, Algorithmic Pricing, and Collusion' (AER)
Q-learning pricing agents under latency, noise, and asynchronous actions — thesis, simulations, and reproducible results.
Simulation of RL agents playing in repeated auction games.
Does LLM price collusion survive replacing an agent? Apart Research AI Collusion Sprint, project 1.2.
Measured costs of anti-collusion levers for LLM market agents: code, 268 runs, pre-registration log (Apart AI Collusion Research Sprint, Oct 2026)
Algorithmic collusion in market making.
A Gymnasium-style suite of economics reinforcement-learning environments with closed-form equilibrium benchmarks.
A study of whether LLM agents shift collusion from chat to bidding when told they are being monitored, comparing chat-based, behavioural, and combined detection methods.
Learned coalition deception and defensive adaptation in a social-deduction benchmark
Do independent Q-learning agents tacitly collude in a uniform-price electricity auction? Three behavioural tests on 50-seed sweeps: high prices yes, collusion not proven.
To associate your repository with the algorithmic-collusion topic, visit your repo's landing page and select "manage topics."