Reproduce “AI agents in Algorithmic Electricity Markets: On the Emergence of Tacit Collusion” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jakub Seredyński, Georgios Tsaousoglou
Autonomous AI agents operating in algorithmic electricity markets can spontaneously learn to engage in tacit collusion, a phenomenon identified through new multi-dimensional criteria, with certain market characteristics like grid constraints and demand levels amplifying this non-competitive behavior. Researchers from the Technical University of Denmark used a Multi-Agent Reinforcement Learning framework to model agent behavior.
Key insights
- Autonomous learning-based AI agents can develop and sustain tacitly collusive strategies in electricity markets without explicit design or instruction to collude.
- Multi-dimensional behavioral criteria provide a more robust and nuanced method for identifying tacit collusion, differentiating it from market power derived purely from structural market features.
- Certain market characteristics, particularly binding transmission constraints, higher demand, and lower cost heterogeneity, significantly enhance the likelihood of tacit collusion among AI agents.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.