Say I have two sets of variables $x$ and $y$ of equal size. $x$'s have a lower bound $x_{min}<0$, and $y$'s have a lower bound $0$.
Is there a linear way to constrain that $x_i\geq0$ if the solution has $y_i>0$, and $x_i\geq x_{min,i}$ if the solution has $y_i=0$.
A few other points that might be useful
- $x$ and $y$ are of length ~100
- $x$ and $y$ both should be integers, but their values are large enough to be treated as continuous, and rounding the solution doesn't hurt too much.
- The objective function is, for a given scalar $t$ and vector $y_0$ $$min\,|\Sigma(x+y)-t| + \Sigma|y-y_0|$$
