Changed math functions to numpy version to allow function vectorization

[belzebuu]

I encountered errors (TypeError: only 0-dimensional arrays can be converted to Python scalars) with vectorized use of the single valued functions, that is, when instead of passing a single number (scalar) as argument, the user gives an array of numbers. In fact, the functions expect an array of values already but in this case the vectorized form has as input a two dimensional array. An example of use that returns the error is passing the matrices from numpy.meshgrid to make 3D plots of the Ackley function.

import numpy as np
from landscapes.single_objective import rosenbrock, booth, ackley

func = ackley

x = np.linspace(-5, 5, 400)
y = np.linspace(-5, 5, 400)
X, Y = np.meshgrid(x, y)
Z = func(np.array([X, Y]))

The reason for this issue seems to be that functions from Python's standard math library (like exp, sqrt, cos), which landscapes uses, are not designed to work on arrays; they only operate on single, scalar values. When we pass the X and Y matrices from meshgrid to the function, we are trying to apply these scalar functions to entire arrays, which causes the TypeError. The solution is to use the equivalent functions from the NumPy library (np.exp, np.sqrt, np.cos), which are specifically designed to operate element-wise on arrays.

Since I needed it, I made some minimal changes myself. I make them available here. Of course, there can be many good reasons for not wanting the merge. Feel free to reject it.