Using sklearn models in mixed linear programming problem?

I have a linear programming problem in which the result from a classifier is used as a variable (binary variable). Only problem is that the classifier itself uses variables in the linear programming problem. Anyone have done something like this before and have some advice on how to proceed? The reason why I do it this way is that I guess I will find the optimal solution faster than if I do it iteratively with some sort of genetic algorithm. …

datascience linear mixed programming sklearn