1: \begin{abstract}

2: In this article we develop a general theory of exact parametric penalty functions for constrained optimization

3: problems. The main advantage of the method of parametric penalty functions is the fact that a parametric penalty

4: function can be both smooth and exact unlike the standard (i.e. non-parametric) exact penalty functions that are always

5: nonsmooth. We obtain several necessary and/or sufficient conditions for the exactness of parametric penalty functions,

6: and for the zero duality gap property to hold true for these functions. We also prove some convergence results for 

7: the method of parametric penalty functions, and derive necessary and sufficient conditions for a parametric penalty

8: function to not have any stationary points outside the set of feasible points of the constrained optimization

9: problem under consideration. In the second part of the paper, we apply the general theory of exact parametric penalty

10: functions to a class of parametric penalty functions introduced by Huyer and Neumaier, and to smoothing

11: approximations of nonsmooth exact penalty functions. The general approach adopted in this article allowed us to unify

12: and significantly sharpen many existing results on parametric penalty functions.

13: \end{abstract}

14: