Subgame consistent solutions for cooperative stochastic dynamic games (Q983725): Difference between revisions

The authors study an \(n\)-person stochastic game, whose state dynamics is given by the equation \[ x_{k+1}= f_k(x_k; u^1_k,\dots, u^n_k)+ \theta_k, \] where \(k= 1,\dots, T\), \(x_1\equiv x^*\), \(u^i_k\in \mathbb{R}^{m_i}\) is a control vector of Player \(i\), \(x_k\in X\subseteq\mathbb{R}^m\) is a state, and \(\theta_k\) is a set of statistically independent random variables. The authors prove that a subgame consistency guarantees dynamical stability. A cooperative solution is subgame consistent if an extension of the solution policy to a subgame starting at a later period with any feasible state induced by prior optimal behaviour remains optimal. The authors illustrate their result by an example of resource extraction.