1: \begin{abstract}

2: In optimal control there may be bad oscillatory phenomena. For instance, this is the case for Fuller's phenomenon \cite{fuller} where the optimal control is bang-bang with an infinite number of switchings in  finite time. 

3: %In problems with state constraints of high order, there are examples in which the optimal  trajectory touches the boundary of the constraint at an infinite number of points over a finite time interval \cite{robbins}. Likewise, 

4: In the framework of hybrid systems, there are optimization problems where the discrete part of the dynamics switches infinitely many times in finite time \cite{zeno-numpb} (and this is known as Zeno's phenomenon.) 

5: 

6: In this paper we provide a technique to regularize this kind of phenomena. Namely, we define a BV regularization of a general optimal control problem and show that the solution of the perturbed problem is  quasi-optimal for the reference problem  in the sense that the loss of optimality is small. We  apply our results %to problems with state constraints and 

7: to hybrid systems and we estimate the decay of the error, when the total variation of the control grows.%, in terms of the sequence of switching times. Switching times have been shown to converge exponentially for  a robust class of problems \cite{kupka}. %Nevertheless, the question whether a slower rate of convergence is typical is still open.

8: \end{abstract}

9: