In this paper, we approximate locally Lipschitz functions with a quadratic function. The search direction is achieved by minimization of this quadratic function. This quadratic model is constructed such that the computed search direction will be descent for the objective function.We apply the line search method along this direction. To update the quadratic model, we use the Quasi-Newton method. Next, we present globally convergent of the minimization algorithm. Finally, the proposed algorithm is implemented in MATLAB environment on the some nonsmooth optimization problems and compared with famous algorithms.
