1: \begin{abstract}

2:   In this letter, we study the proximal gradient dynamics. This

3:   recently-proposed continuous-time dynamics solves optimization problems

4:   whose cost functions are separable into a nonsmooth convex and a smooth

5:   component.

6:   %%

7:   First, we show that the cost function decreases monotonically along the

8:   trajectories of the proximal gradient dynamics.

9:   %

10:   We then introduce a new condition that guarantees exponential convergence

11:   of the cost function to its optimal value, and show that this condition

12:   implies the proximal Polyak-\L{ojasiewicz} condition. 

13:   %%

14:   We also show that the proximal Polyak-\L{ojasiewicz} condition guarantees exponential convergence of the cost function.

15:   %%

16:   Moreover, we extend these results to time-varying optimization problems,

17:   providing bounds for equilibrium tracking.

18:   %%

19:   Finally, we discuss applications of these findings, including the LASSO

20:   problem, quadratic optimization with polytopic constraints, and certain matrix based problems.

21: \end{abstract}

22: