This paper introduces competitive mediator games, a new framework for analyzing equilibria in markets, particularly for autonomous routing and driving (ARAD) services. These games generalize correlated equilibria and are shown to be limits of algorithmic no-regret learning dynamics. The research demonstrates that in anonymous congestion games with market-share maximizing mediators, all competitive mediator equilibria result in monopolies when one mediator is weakly preferred by all users. The authors also provide an overview of ARAD markets and discuss future mechanism design. AI
IMPACT Introduces a new game theory framework for analyzing autonomous systems and market competition.
RANK_REASON Academic paper published on arXiv. [lever_c_demoted from research: ic=1 ai=0.7]
Read on arXiv cs.MA (Multiagent) →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
