I want to optimize a function with scipy.optimize.minimize, that calls a function that uses at every call a new random set by np.random.normal(m,s,N). Therefore at every call the function results in slightly different results. result.success is finally False.

If I track the optimization calls inside my function I see, that there is a best solution. Is it possible to keep the best found solution and get it from scipy.otpimize.minimize?