Duality-based a posteriori error estimates for some approximation schemes for optimal investment problems (Q2212323)

Numerical approximation of a class of optimal control problems for diffusion processes arising in financial applications is studied. The associated value function of such problems can be characterized as the solution of the Hamilton-Jacobi-Bellman partial differential equation, for which the explicit solution is usually not available. Therefore, a numerical approximation of the solution appears to be important from the practical point of view. The authors concentrate their attention on computable error bounds for the solution. For a suitable class of convex optimal control problems, a posteriori error bounds are obtained using the numerical approximation of a dual problem. Advantages of the proposed error bounds are discussed. Convergence properties of the proposed procedure are demonstrated on numerical test examples. Perspectives of further research are outlined in the conclusion section.