1: \begin{abstract}

2:   We propose an algorithm for solving bound-constrained mathematical programs

3:   with complementarity constraints on the variables. Each iteration of the algorithm involves

4:   solving a linear program with complementarity constraints in order to obtain

5:   an estimate of the active set. The algorithm enforces

6:   descent on the objective function to promote global convergence to

7:   B-stationary points. We provide a convergence analysis and preliminary

8:   numerical results on a range of test problems.

9:   We also study the effect of fixing the active

10:   constraints in a bound-constrained quadratic program that can be solved on

11:   each iteration in order to obtain fast convergence. 

12: \end{abstract}

13: