Numerical differentiation for the second order derivatives of functions of two variables (Q2468138)

For computing the second-order derivatives of functions with two variables, the authors propose a numerical differentiation technique based on noisy scattered data. This is an ill-posed problem, that is, the small errors in the measurement of the function may lead to large errors in its computed derivatives. The problem is formulated as a regularized optimization problem and so the Tikhonov regularization method is applied. The existence and uniqueness of the solution to this problem are proved and a constructive scheme for the approximate regularized solution are described. For this purpose, the bi-harmonic Green's function is used. A convergence estimate of the regularized solution to the exact solution, is derived. Numerical examples are performed and the results show that the proposed method is efficient and promising.