1: \begin{abstract}

2:     We re-introduce a derivative-free subspace optimization framework originating from Chapter 5 of the

3:     Ph.D.~thesis [Z. Zhang, \textit{On Derivative-Free Optimization Methods}, Ph.D.~thesis, Chinese Academy

4:     of Sciences, Beijing, 2012] of the author under the supervision of Ya-xiang Yuan.

5:     At each iteration, the framework defines a (low-dimensional) subspace based on an approximate gradient, and

6:     then solves a subproblem in this subspace to generate a new iterate.

7:     We sketch the global convergence and worst-case complexity analysis of the framework,

8:     elaborate on its implementation, and present some numerical results on solving problems with

9:     dimension as high as $10^4$ using only inaccurate function values.

10:     %

11:     %

12:     %

13: \end{abstract}

14: