1: \begin{abstract}

2: In this paper, we carry out a unified study for $L_1$ over $L_2$  sparsity promoting models, which are widely used in the regime of coherent dictionaries for recovering sparse nonnegative/arbitrary signals. First, we provide a unified theoretical analysis on  the  existence of the global solutions of the constrained and the unconstrained $L_{1}/L_{2}$ models.

3: Second, we analyze the sparse property of any local minimizer of these $L_{1}/L_{2}$ models which serves as a certificate to rule out

4: the nonlocal-minimizer stationary solutions.

5: Third, we derive an analytical solution for the proximal operator of the $L_{1} / L_{2}$ with nonnegative constraint.

6:  Equipped with this, we apply the alternating direction method of multipliers   to  the unconstrained model with nonnegative constraint in a particular splitting way,

7: referred to as ADMM$_p^+$. We establish its global convergence to a  d-stationary solution (sharpest stationary) without

8:  the Kurdyka-\L ojasiewicz assumption.

9: Extensive numerical simulations  confirm the superior  of ADMM$_p^+$ over

10:  the state-of-the-art methods in sparse recovery. In particular, ADMM$_p^+$ reduces computational time by about $95\%\sim99\%$ while achieving a much higher accuracy than the commonly used  scaled gradient projection method for the wavelength misalignment problem.

11: \keywords{Sparse recovery \and fractional programming \and coherent dictionary \and d-stationarity}

12: % \PACS{PACS code1 \and PACS code2 \and more}

13: \subclass{MSC 90C26 \and MSC    90C90 \and 49N45}

14: \end{abstract}