A Priori and A Posteriori Error Analysis in Chemistry
Convergence analysis; Error estimates; Guaranteed accuracy; Refinement
For a numerical discretization chosen to approximate the solution of a given problem or for an algorithm used to solve the discrete problem resulting from the previous discretization, a priori analysis explains how the method behaves and to which extent the numerical solution that is produced from the discretization/algorithm is close to the exact one. It also allows to compare the numerical method of interest with another one. With a priori analysis though, there is no definite certainty that a given computation provides a good enough approximation. It is only when the number of degrees of freedom and the complexity of the computation is large enough that the convergence of the numerical method can be guaranteedly achieved. On the contrary, a posteriori analysis provides bounds on the error on the solution or the output coming out of the simulation. The concept of a posteriori analysis can even...
- 1.Boulton, L., Boussa, N., Lewin, M.: Generalized Weyl theorem and spectral pollution in the Galerkin method. http://arxiv.org/pdf/1011.3634v2
- 5.Cancès, E., Ehrlacher, V., Maday, Y.: Periodic Schrdinger operators with local defects and spectral pollution, arXiv:1111.3892Google Scholar
- 8.Cancès, E., LeBris, C., Maday, Y.: Méthodes Mathématiques en chimie quantique: une Introduction (in French). Mathématiques and Applications (Berlin), vol. 53. Springer, Berlin (2006)Google Scholar
- 9.Conte, D., Lubich, C.: An error analysis of the multi-configuration time-dependent Hartree method of quantum dynamics. ESAIM M2AN 44, 759–780 (2010)Google Scholar
- 13.Kobus, J., Quiney, H.M., Wilson, S.: A comparison of finite difference and finite basis set Hartree-Fock calculations for the N2 molecule with finite nuclei. J. Phys. B Atomic Mol. Opt. Phys. 34, 10 (2001)Google Scholar