Approximation methods for the finite moment problem (Q1315219)

The finite Hausdorff moment problem is considered within the context of Sobolev spaces of order 0, 1, or 2. In each case the normal solution of the finite moment problem is approximated by orthogonal functions (polynomials in the case of Sobolev spaces of order 0 or 1) which are computed by a computationally efficient algorithm. The problem is considered both for exact data and noisy data. In the case of noisy data the problem is ill-posed and an a posteriori method is suggested to select the normal solution which minimizes the uniform norm of the recovery error. The method is illustrated on a number of standard test problems.