α-Rank: Multi-Agent Evaluation by Evolution

Scientific reports, pp. 99372019.

Cited by: 0|Bibtex|Views55|Links
EI WOS

Abstract:

We introduce α-Rank, a principled evolutionary dynamics methodology, for the evaluation and ranking of agents in large-scale multi-agent interactions, grounded in a novel dynamical game-theoretic solution concept called Markov-Conley chains (MCCs). The approach leverages continuous-time and discrete-time evolutionary dynamical systems app...More

Code:

Data:

Your rating :
0

 

Tags
Comments