how to dynamically set the number of variables in SciPy's optimize.minimize

Question

I have this code work fine to optimize multiple variables.

def f(params):

    a, b, c = params 
    return a**2 + b**2 + c**2

initial_guess = [1, 1, 1]
result = optimize.minimize(f, initial_guess)
if result.success:
    fitted_params = result.x
    print(fitted_params)
else:
    raise ValueError(result.message)

Is it possible to modify the code so that the number of parameters can be dynamically passed into f? For example, I'd like to have something like this.

def f(params, paramsNumber):
    if paramsNumber == 3:
        a, b, c = params 
        return a**2 + b**2 + c**2
    elif paramsNumber == 2:
        a, b  = params 
        return a**2 + b**2 

if paramsNumber == 3:
    initial_guess = [1, 1, 1]
elif paramsNumber == 2:
    initial_guess = [1, 1]
result = optimize.minimize(f, initial_guess)
if result.success:
    fitted_params = result.x
    print(fitted_params)
else:
    raise ValueError(result.message)

Thanks a lot!

Answer 1

You can iterate over the params in your objective function, and you can use some general initialization, for your example I simply used numpy ones, that create an array with the given number of ones.

import numpy as np
from scipy import optimize
def f(params):
    return sum(a**2 for a in params)
def find_sol(paramsNumber):
    initial_guess = np.ones(paramsNumber)
    result = optimize.minimize(f, initial_guess)
    if result.success:
        return result.x
    raise ValueError(result.message)

find_sol(4)

array([-3.15727582e-08, -3.15727581e-08, -2.06550776e-09, -2.06550765e-09])