Sparse nonnegative solution of underdetermined linear system via non-convex fraction function penalty

Angang Cui,Haiyang Li,Meng Wen,Jigen Peng

2017-01-01

Abstract:In this paper, a non-convex fraction function is studied in sparse nonnegative solution of an underdetermined linear system. Firstly, we discuss the equivalence of $(FP_{a}^{geq})$ and $(P_{0}^{geq})$, and proved that there exists some constant $hat{a}u003e0$ such that the optimal solution to the problem $(FP_{hat{a}}^{geq})$ also solves $(P_{0}^{geq})$. Furthermore, we prove that there corresponds a constant $a^{*}u003e0$ such that, whenever $au003ea^{*}$, every solution to the problem $(FP_{a}^{geq})$ also solves $(P_{0}^{geq})$. Secondly, a nonnegative iterative thresholding algorithm is proposed to solve the regularization problem $(FP_{a,lambda}^{geq})$ for all $au003e0$. Finally, some numerical experiments on sparse nonnegative vector recovery problems show that our method performs effective in finding a sparse nonnegative solution.