Random Graphs Estimation using Q-Learning
Mina Babahaji, Stéphane Blouin, Walter Lúcia, Mohammad Mehdi Asadi, Hamid Mahboubi, Amir G. Aghdam · 2021
To correctly estimate a random graph it is important to be able to estimate, in a distributed fashion, the probability matrix (characterizing the probability of the existence of the graph’s edges) and the graph connectivity. In this paper, first, by leveraging Q-learning arguments, two different solutions to the probability estimation problem are proposed. Then, a method for the estimation of the algebraic connectivity is given. The accuracy of the proposed methods are verified by simulation for an underwater sensor network.