Figure 17 | Scientific Reports

Figure 17

From: α-Rank: Multi-Agent Evaluation by Evolution

Figure 17

A retrospective look on the paper contributions. We introduced a general descriptive multi-agent evaluation method, called \(\alpha \)-Rank, which is practical in the sense that it is easily applicable in complex game-theoretic settings, and theoretically-grounded in a solution concept called Markov-Conley chains (MCCs). \(\alpha \)-Rank has a strong theoretical and specifically evolutionary interpretation; the overarching perspective considers a chain of models of increasing complexity, with a discrete-time macro-dynamics model on one end, continuous-time micro-dynamics on the other end, and MCCs as the link in between. We provided both scalability properties and theoretical guarantees for the overall ranking methodology.

Back to article page