On the Hamilton-Jacobi-Bellman equations (Q1077371)

The author considers optimal stochastic control problems and the associated Hamilton-Jacobi-Bellman equations. The heuristic argument showing that the minimal cost function satisfies the H-J-B equation is given. Then the author shows that the minimal cost function, u, is the maximum element of the set of all sub-solutions v satisfying: \(A_{\alpha}v\leq f_{\alpha}\) in \(R^ n\), for all \(\alpha\in A\) \((A_{\alpha}\) elliptic operator). Continuity and regularity results are obtained for u, which shows u satisfies the H-J-B in a \(W^{2,\infty}\)- sense. Viscosity solutions for general second order PDE are defined.