Improve the code.

- Add a docstring.
- Add type hinting.
- Change algorithm from steepest ascent to simple hill climbing.

hill_climbing.py

@@ -1,53 +1,66 @@
+import copy
 from typing import Tuple
 
 import numpy as np
 import numpy.typing as npt
-import copy
 
-from common import plot_plan, read_data
+from common import indexes_to_cities, plot_plan, read_data
 
-original_perm = None
 
 def hill_climbing(distances: npt.NDArray) -> Tuple[float, npt.NDArray]:
-    size = len(distances)
+    """A simple hill climbing algorithm.
-    perm = np.arange(size)
-    np.random.shuffle(perm)
 
-    print(perm)
+    The algorithm starts on a random permutation and attempts to improve
-    global original_perm
+    the circuit by trying to switch neighboring elements. Each iteration
-    original_perm = copy.deepcopy(perm)
+    tries to switch adjacent neighbors and sees which one yields the largest
-    current_route: float = np.sum([distances[perm[i - 1], perm[i]] for i in range(size)])
+    improvement.
-    found_improvement = True
-    while found_improvement:
-        found_improvement = False
-        tmp_improvement: float = current_route
-        improvement_index: int = -1
-        for i in range(size):
-            perm[i - 1], perm[i] = perm[i], perm[i - 1]
-            tmp_route: float = np.sum([distances[perm[i - 1], perm[i]] for i in range(size)])
-            if tmp_route < tmp_improvement:
-                tmp_improvement = tmp_route
-                improvement_index = i
-                found_improvement = True
 
-            perm[i - 1], perm[i] = perm[i], perm[i - 1]
+    Args:
+        distances npt.NDArray: A matrix containing the distances between cities.
 
-        if found_improvement:
+    Returns:
-            current_route = tmp_improvement
+        Tuple[float, npt.NDArray] A tuple containing the distance of the
-            perm[improvement_index - 1], perm[improvement_index] = (
+        solution and the solution itself.
-                perm[improvement_index],
+
-                perm[improvement_index - 1],
+    """
+
+    size: int = len(distances)  # The size of the permutation array
+    perm: npt.NDArray = np.arange(size)  # Create an array from 0..size
+
+    np.random.shuffle(perm)  # Get random permutation
+
+    # Get the distance of the random permutation
+    current_distance: float = np.sum(
+        [distances[perm[i - 1], perm[i]] for i in range(size)]
     )
 
-    print(perm)
+    found_improvement: bool = True
 
-    return (current_route, perm)
+    while found_improvement:
+
+        found_improvement = False  # Assume we haven't found an improvement
+        tmp_distance: float = current_distance
+
+        # Try to find an improvement
+        for i in range(size):
+            perm[[i - 1, i]] = perm[[i, i - 1]]  # Swap i - 1 and i
+
+            tmp_distance: float = np.sum(
+                [distances[perm[i - 1], perm[i]] for i in range(size)]
+            )
+
+            if tmp_distance < current_distance:
+                current_distance = tmp_distance
+                found_improvement = True
+                break
+
+            perm[[i - 1, i]] = perm[[i, i - 1]]  # Swap back i - 1 and i
+
+    return (current_distance, perm)
 
 
 if __name__ == "__main__":
     cities, data = read_data("./european_cities.csv")
-    distance, perm = hill_climbing(data[:10,:10])
+    distance, perm = hill_climbing(data[:10, :10])
-
-    plot_plan(list(map(lambda i: cities[i], list(perm))))
-    plot_plan(list(map(lambda i: cities[i], list(original_perm))))
 
+    plot_plan(indexes_to_cities(perm, cities))