From bcf22269e8fde580db619ff89ff0b32da1133dc5 Mon Sep 17 00:00:00 2001
From: Peter Vacho <p_vacho@utb.cz>
Date: Tue, 11 Mar 2025 01:02:41 +0100
Subject: [PATCH] Add necessary utils

---
 src/function.py  |  42 +++++++++++++++++
 src/functions.py | 116 +++++++++++++++++++++++++++++++++++++++++++++++
 src/types.py     |   4 ++
 3 files changed, 162 insertions(+)
 create mode 100644 src/function.py
 create mode 100644 src/functions.py
 create mode 100644 src/types.py

diff --git a/src/function.py b/src/function.py
new file mode 100644
index 0000000..5b9106b
--- /dev/null
+++ b/src/function.py
@@ -0,0 +1,42 @@
+from collections.abc import Callable
+from dataclasses import dataclass
+from typing import NamedTuple
+
+import numpy as np
+
+from src.types import INPUT_VECTOR
+
+
+class Interval(NamedTuple):
+    """Interval determining a number range (both inclusive)."""
+
+    min: float
+    max: float
+
+    def random_point(self, dimensions: int, rng: np.random.Generator | None = None) -> INPUT_VECTOR:
+        """Generate an N-dimensional random point, where all numbers lie within this interval.
+
+        Args:
+            dimensions: The amount of dimensions this point should have.
+            rng: Random generator instance (None for a new rng).
+        """
+        if rng is None:
+            rng = np.random.default_rng()
+
+        return rng.uniform(self.min, self.max, size=dimensions)
+
+
+@dataclass
+class Function:
+    """Class representing an N-dimensional function.
+
+    This function can work with arbitrary number of dimensions, where each input dimension is restricted
+    within the specified definition interval.
+    """
+
+    function: Callable[[INPUT_VECTOR], float]
+    definition_interval: Interval
+
+    def __call__(self, params: INPUT_VECTOR) -> float:
+        """Evaluate the function."""
+        return self.function(params)
diff --git a/src/functions.py b/src/functions.py
new file mode 100644
index 0000000..fe6e1b2
--- /dev/null
+++ b/src/functions.py
@@ -0,0 +1,116 @@
+"""Various N-dimensional well-known optimization functions.
+
+All of the functions here work with N (# of dimensions) ranging from 1 to infinity.
+"""
+
+import numpy as np
+
+from src.function import Function, Interval
+from src.types import INPUT_VECTOR
+
+
+def _sphere_function(vector: INPUT_VECTOR) -> float:
+    """Compute the sphere function.
+
+    Arguments:
+        vector: An n-dimensional input vector.
+
+    Returns:
+        float: The function value at the given input vector.
+
+    See: https://www.sfu.ca/~ssurjano/spheref.html
+    """
+    return np.sum(vector**2)
+
+
+sphere_function = Function(_sphere_function, Interval(-10, 10))
+
+
+def _zakharov_function(vector: INPUT_VECTOR) -> float:
+    """Compute the Zakharov function.
+
+    Arguments:
+        vector: An n-dimensional input vector.
+
+    Returns:
+        float: The function value at the given input vector.
+
+    See: https://www.sfu.ca/~ssurjano/zakharov.html
+    """
+    d = vector.shape[0]
+
+    # First term: sum(x_i^2)
+    term_1 = np.sum(vector**2)
+
+    # Second term: (sum(0.5 * i * x_i))^2
+    term_2 = np.sum(0.5 * np.arange(1, d + 1) * vector) ** 2
+
+    # Third term: (sum(0.5 * i * x_i))^4
+    term_3 = np.sum(0.5 * np.arange(1, d + 1) * vector) ** 4
+
+    return term_1 + term_2 + term_3
+
+
+zakharov_function = Function(_zakharov_function, Interval(-10, 10))
+
+
+def _rosenbrock_function(vector: INPUT_VECTOR) -> float:
+    """Compute the Rosenbrock function.
+
+    Arguments:
+        vector: An n-dimensional input vector.
+
+    Returns:
+        float: The function value at the given input vector.
+
+    See: https://www.sfu.ca/~ssurjano/rosenbrockf.html
+    """
+    return np.sum(100 * (vector[:-1] ** 2 - vector[1:]) ** 2 + (1 - vector[:-1]) ** 2)
+
+
+rosenbrock_function = Function(_rosenbrock_function, Interval(-10, 10))
+
+
+def _schwefel_function(vector: INPUT_VECTOR) -> float:
+    """Compute the Schwefel function.
+
+    Arguments:
+        vector: An n-dimensional input vector.
+
+    Returns:
+        float: The function value at the given input vector.
+
+    See: https://www.sfu.ca/~ssurjano/schwefel.html
+    """
+    d = vector.shape[0]
+
+    return 418.9829 * d - np.sum(vector * np.sin(np.sqrt(np.abs(vector))))
+
+
+schwefel_function = Function(_schwefel_function, Interval(-500, 500))
+
+
+def _styblinski_tang_function(vector: INPUT_VECTOR) -> float:
+    """Compute the Styblinski-Tang function.
+
+    Arguments:
+        vector: An n-dimensional input vector.
+
+    Returns:
+        float: The function value at the given input vector.
+
+    See: https://www.sfu.ca/~ssurjano/stybtang.html
+    """
+    return 0.5 * np.sum(vector**4 - 16 * vector**2 + 5 * vector)
+
+
+styblinski_tang_function = Function(_styblinski_tang_function, Interval(-5, 5))
+
+
+available_functions = {
+    "Sphere": sphere_function,
+    "Zakharov": zakharov_function,
+    "Rosenbrock": rosenbrock_function,
+    "Schwefel": schwefel_function,
+    "Styblinski-Tang": styblinski_tang_function,
+}
diff --git a/src/types.py b/src/types.py
new file mode 100644
index 0000000..a868754
--- /dev/null
+++ b/src/types.py
@@ -0,0 +1,4 @@
+import numpy as np
+import numpy.typing as npt
+
+type INPUT_VECTOR = npt.NDArray[np.float64]