A learning approach to the two person decentralized team problem with incomplete information

A learning approach to the two person decentralized team problem with incomplete information and with a 2X2 payoff matrix is considered. It is shown that if the payoff matrix is unimodal, there exists a proper choice of parameters of the learning algorithm that will ensure asymptotically an expected payoff as close to maximum payoff as desired. It is shown that a team problem with multimodal payoff matrix gives rise to interesting class of open problems.