Quality of Non-Convergent Best Response Processes in Multi-Agent Systems through Sink Equilibrium
Rohit Konda, Rahul Chandan, Jason Marden
TL;DR
This work addresses the challenge of analyzing multi-agent systems when self-interested best-response processes may not converge to Nash equilibria. It adopts sink equilibria as attractors of the best-response dynamics and introduces price of sinking (PoSE) to quantify welfare under these equilibria. By integrating a novel misalignment parameter (beta) with a relaxed smoothness framework, the authors derive nontrivial lower bounds on PoSE under arithmetic and geometric misalignment, illustrating that non-convergent BR processes can still yield bounded efficiency losses. They further show that sink equilibria induced by better responses guarantee the existence of high-welfare actions within the sink's support, and discuss implications for distributed applications and future research directions, including tighter bounds and broader sink-inducing dynamics.
Abstract
Examining the behavior of multi-agent systems is vitally important to many emerging distributed applications - game theory has emerged as a powerful tool set in which to do so. The main approach of game-theoretic techniques is to model agents as players in a game, and predict the emergent behavior through the relevant Nash equilibrium. The virtue from this viewpoint is that by assuming that self-interested decision-making processes lead to Nash equilibrium, system behavior can then be captured by Nash equilibrium without studying the decision-making processes explicitly. This approach has seen success in a wide variety of domains, such as sensor coverage, traffic networks, auctions, and network coordination. However, in many other problem settings, Nash equilibrium are not necessarily guaranteed to exist or emerge from self-interested processes. Thus the main focus of the paper is on the study of sink equilibrium, which are defined as the attractors of these decision-making processes. By classifying system outcomes through a global objective function, we can analyze the resulting approximation guarantees that sink equilibrium have for a given game. Our main result is an approximation guarantee on the sink equilibrium through defining an introduced metric of misalignment, which captures how uniform agents are in their self-interested decision making. Overall, sink equilibrium are naturally occurring in many multi-agent contexts, and we display our results on their quality with respect to two practical problem settings.