题 目：Efficient Preconditioning for Inverse Source Problems in Time Fractional Diffusion Equations
报 告 人：柯日焕 博士后 (邀请人：彭小飞 )
时 间：2020-01-06 15:00--16：00
The problem of identifying unknown source terms for the time-fractional diffusion equations is ill-posed. In this talk, we consider a quasi-boundary value regularization method for recovering a space-dependent source term from the final observation data. In particular, an efficient preconditioning technique motivated by the block structure of the linear system is proposed. Spectrum analysis shows that the eigenvalues of the preconditioned system are clustered around one, and hence fast convergence of the method is guaranteed. An L-Curve type method is developed for selecting the regularization parameters. Numerical experiments are presented to demonstrate the effectiveness of the proposed methods.