Regularization by truncated Cholesky factorization: a comparison of four different approaches (Q883332)

Due to the principle of regularization by restricting the number of degrees of freedom, truncating the Cholesky factorization of a symmetric positive definite matrix can be expected to have a stabilizing effect. Based on this idea, the author considers four different approaches for regularizing ill-posed linear operator equations. Convergence in the noise free case and in the situation of noisy data is analyzed. Numerical tests illustrate the theoretical results.