To run on test data (which is in testdata.h), do g++ -I. -O3 test.cpp -o test, and then ./test.
To run on your own data, include fgfunction.h, put your data in a FGData structure,
create a FGFunction object, and call runBFGS() on that. The dependencies LBFGSpp, fgdata.h, and objgradfun.h
should be present in the same directory, and Eigen should be in your include path.
Example usage:
#include "fgfunction.cpp"
int main() {
// this is your data
FGData data = {
.E = &equality_constraints,
.L = &lower_bounds,
.U = &upper_bounds,
.w = w,
.f = f
};
// init_x is your starting point for the minimsation
// (which is X0 in postProcessingMe.m)
FGFunction f(init_x, &data);
// run the BFGS
// note that passing init_x will overwrite it with the result,
// so pass a copy if you want to preserve it
double out = f.runBFGS(init_x);
// print out the minimal objective function value and the result
printf("minimisation = %lf\n", out);
for (double i : test_rawX)
printf("%lf ", i);
printf("\ndone\n");
}