[Python] Plugging SciPy's Nelder-Mead result back into the function does not return reported...

When I take the x reported by the optimization and plug it back into my function optimizee, I don't get the value outputted by the optimization: -749.260. Rather I get -637.65, which is not as good...

I'm using SciPy's Nelder-Mead to optimize a nasty function I have put together:

result = minimize(optimizee, x0, method='Nelder-Mead', bounds=(
(0, 250),
(0, 250),
(0, 1),
(0, 1),
(0, 1),
(0, 1),
(0, 1),
(0, 1),
(0, 1),
(0, 1),
(0, 1),
(0, 1),
), options={'xatol': 1e-2, 'maxiter': 50000})


By raising maxiter, I get it to to converge and successfully return:

message: Optimization terminated successfully.
success: True
status: 0
fun: -749.2601549652912
x: [ 5.536e-03 2.500e+02 6.156e-04 0.000e+00 1.033e-04
2.533e-06 1.000e+00 5.640e-01 8.739e-01 3.828e-03
9.999e-01 1.000e+00]
nit: 1811
nfev: 2589
final_simplex: (array([[ 5.536e-03, 2.500e+02, ..., 9.999e-01,
1.000e+00],
[ 5.557e-03, 2.500e+02, ..., 9.999e-01,
1.000e+00],
...,
[ 5.788e-03, 2.500e+02, ..., 9.999e-01,
1.000e+00],
[ 5.782e-03, 2.500e+02, ..., 1.000e+00,
1.000e+00]], shape=(13, 12)), array([-7.493e+02, -7.493e+02, ..., -7.493e+02, -7.493e+02],
shape=(13,)))

Continue reading...