Quadratic programming and the single-controller stochastic game (Q1096555)

The author considers an approximately constructed quadratic program, the optimal solution of which provides a stationary Nash-equilibrium point of a two-person, general-sum, single-controller stochastic game. He considers stochastic games with both limiting average and discounted payoff criteria. Subsequently, it is shown that for the latter, the converse statement also holds i.e. every stationary equilibrium point provides an optimal solution to the quadratic program. It is observed that the results include as special cases non-quadratic linear programming formulations of bi-matrix games, matrix games, Markovian decision processes and single controller zero-sum stochastic games.