L1 norm in objective function


1 comment

  • Yuriy Zinchenko

    Yes, absolutely.  Although we do not have an explicit '\ell-1' construct to accommodate the above norm, we have the absolute value helper function, e.g., 


    that can be used to get the desired expression. Also note that if you aim to simply minimize the L1 norm expression, you may as well build it yourself as a sequence of convex inequalities that involve auxiliary variables, e.g.,

    min ||x||_1

    is equivalent to

    min t


    x_i <= y_i,

    -x_i <= y_i,

    \sum_i y_i == t

    Comment actions Permalink

Please sign in to leave a comment.

Powered by Zendesk