Second-order cone reformulation and the price of anarchy of a robust nash-cournot game

We study an n-person Nash-Cournot game with incomplete information, in which the opponents' strategies are only known in a perturbed set and the players try to minimize their worst-case costs, which can vary due to data uncertainty. We show that in several interesting cases, this game can be reformulated as second-order cone optimization problems. We also derive a bound of the price of anarchy for this game, which is a bound on the ratio between the cost at the robust Nash-Cournot equilibria and the cost at the system optima.