1: \begin{abstract}

2: We propose a bundle trust-region algorithm

3: to minimize locally Lipschitz functions which are potentially nonsmooth

4: and nonconvex. We prove global convergence of our method and show by way of an example

5: that the classical convergence argument in trust-region methods based

6: on the Cauchy point fails in the nonsmooth setting.

7: Our method is tested experimentally on three problems in automatic control.

8: 

9: \vskip 8pt

10: \noindent{\bf Keywords.} Bundle $\cdot$ cutting plane $\cdot$  trust-region  $\cdot$ Cauchy point $\cdot$ 

11: global convergence $\cdot$   parametric robustness $\cdot$

12: distance to instability $\cdot$ worst-case $H_\infty$-norm 

13: \end{abstract}

14: