CSIAM Trans. Appl. Math., 2 (2021), pp. 108-130.
Published online: 2021-02
Cited by
- BibTex
- RIS
- TXT
The primal dual fixed point (PDFP) proposed in [7] was designed to solve convex composite optimization problems in imaging and data sciences. The algorithm was shown to have some advantages for simplicity and flexibility for divers applications. In this paper we study two modified schemes in order to accelerate its performance. The first one considered is an inertial variant of PDFP, namely inertial PDFP (iPDFP) and the second one is based on a prediction correction framework proposed in [20], namely Prediction Correction PDFP (PC-PDFP). Convergence analysis on both algorithms is provided. Numerical experiments on sparse signal recovery and CT image reconstruction using TV-$L_2$ model are presented to demonstrate the acceleration of the two proposed algorithms compared to the original PDFP algorithm.
}, issn = {2708-0579}, doi = {https://doi.org/10.4208/csiam-am.2020-0042}, url = {http://global-sci.org/intro/article_detail/csiam-am/18656.html} }The primal dual fixed point (PDFP) proposed in [7] was designed to solve convex composite optimization problems in imaging and data sciences. The algorithm was shown to have some advantages for simplicity and flexibility for divers applications. In this paper we study two modified schemes in order to accelerate its performance. The first one considered is an inertial variant of PDFP, namely inertial PDFP (iPDFP) and the second one is based on a prediction correction framework proposed in [20], namely Prediction Correction PDFP (PC-PDFP). Convergence analysis on both algorithms is provided. Numerical experiments on sparse signal recovery and CT image reconstruction using TV-$L_2$ model are presented to demonstrate the acceleration of the two proposed algorithms compared to the original PDFP algorithm.