A Stochastic Decision Problem
In a team decision problem there are n agents. Agent i observes random variable Y i and, as a function of this observation, takes decision u i from a given set U i of possible decisions. Denoting the decision function by γ i, the problem is to choose (γ 1,...γ n) so as to optimize the expectation of a criterion C(u 1, ..., u n, Z), where Z is a random variable and the joint distribution of Z and the Y i is given . Note that by conditioning one can assume that Z is the n-tuple of all observations Y i. Outside a few special cases, team problems are of high complexity .