The regularized weak functional matching pursuit for linear inverse problems (Q2316685)

This work presents regularized weak functional matching pursuit (RWFMP) for solving linear inverse problems in Hilbert spaces. It extends regularized functional matching pursuit (RFMP) developed in the authors' prior work using an idea from weak greedy approximation, and overcomes certain drawbacks of RFMP, e.g., existence of maximizing element and finite-dimensional range. The study includes several theoretical results, including convergence and rates for suitable parameter choice. The algorithm is illustrated with one-dimensional model experiments.