The task is to find the minimum of function with explicitly defined step. I went through methods of scipy.optimize package, which use approximation approach (COBYLA, COBYQA), and didn't find any parameters and options, which can be used to pass the step size. For example: initial_x_1 = np.float16([4.0]) res_opt_x_1 = optimize.minimize(f_x_1, initial_x_1, method='COBYLA', options={'rhobeg': [2.0], 'maxiter': 50, 'disp': False, 'catol': 1e-4}) Is the other numeric methods in scipy.optimize, in which we can set the size of step explicitly? Or, more generally, are there other numeric methods in OTHER Python mathematic libraries, in which we can set the size of step explicitly? Continue reading...
