Hi all,
I'm trying to model a problem but it turned out to be non linear.
A simplified version of the model is written below. Basically it averages the weighted value of all enabled points, provided there are exactly M enabled points.
max sum(i) { enabled[i] * value[i] * weight[i] } / sum(i) { enabled[i] * weight[i] }
s.t. sum (i) enabled[i] = M
- value is a vector of decimal numbers in [0, 1] (precomputed)
- weight is a vector of decimal numbers in [0, 1] (precomputed)
- enabled is a vector of either 0 or 1 (decision variable)
The model is very simple so I'm guessing there probably is a way to linearize it or some workaround I'm not aware of.
Thanks,
Matt