Dynamic Information Sharing and Punishment Strategies

In this article, we study the problem of information sharing among rational self-interested agents as a dynamic game asymmetric information. We assume that imperfectly observe Markov chain, and they are called to decide whether will share their noisy observations or not at each time instant. utilize notion conditional mutual information evaluate being shared agents. The challenges arise due interdependence agents’ structure decision making exhibited. For finite horizon game, prove do have incentive contrast, show cooperation can be sustained in infinite-horizon case by devising appropriate punishment strategies, which defined over beliefs on system state. these strategies closed under best-response mapping optimal choice some subsets state belief simplex. characterize equilibrium regions, uniqueness maximal equilibrium region, devise an algorithm for its approximate computation.