def test_random():
"""Test random method"""
dists = [(0, 1, 3), (0, 2, 5), (0, 3, 1), (0, 4, 7), (1, 3, 6),
(4, 1, 9), (2, 3, 8), (2, 4, 2), (3, 2, 8), (3, 4, 4)]
problem = TSPOpt(5, distances=dists)
rand = problem.random()
assert (len(rand) == 5 and len(set(rand)) == 5)
def test_adjust_probs_all_zero():
"""Test adjust_probs method when all elements in input vector sum to
zero."""
dists = [(0, 1, 3), (0, 2, 5), (0, 3, 1), (0, 4, 7), (1, 3, 6),
(4, 1, 9), (2, 3, 8), (2, 4, 2), (3, 2, 8), (3, 4, 4)]
problem = TSPOpt(5, distances=dists)
probs = np.zeros(5)
assert np.array_equal(problem.adjust_probs(probs), np.zeros(5))
def test_adjust_probs_non_zero():
"""Test adjust_probs method when all elements in input vector sum to
some value other than zero."""
dists = [(0, 1, 3), (0, 2, 5), (0, 3, 1), (0, 4, 7), (1, 3, 6),
(4, 1, 9), (2, 3, 8), (2, 4, 2), (3, 2, 8), (3, 4, 4)]
problem = TSPOpt(5, distances=dists)
probs = np.array([0.1, 0.2, 0, 0, 0.5])
x = np.array([0.125, 0.25, 0, 0, 0.625])
assert np.array_equal(problem.adjust_probs(probs), x)
def test_reproduce_mut1():
"""Test reproduce method when mutation_prob is 1"""
dists = [(0, 1, 3), (0, 2, 5), (0, 3, 1), (0, 4, 7), (1, 3, 6),
(4, 1, 9), (2, 3, 8), (2, 4, 2), (3, 2, 8), (3, 4, 4)]
problem = TSPOpt(5, distances=dists)
father = np.array([0, 1, 2, 3, 4])
mother = np.array([4, 3, 2, 1, 0])
child = problem.reproduce(father, mother, mutation_prob=1)
assert (len(child) == 5 and len(set(child)) == 5)
def __compute_shortest_path(self, starting_zone_id, zones_ids):
"""
Compute the shortest path within a sequence of geo-localized zones by solving the
Travelling Salesperson Problem (TSP) optimization problem with a genetic algorithm.
"""
zones = [starting_zone_id] + zones_ids
coords_list = []
# Extract zones coordinates
grid_h = self.sim_input.grid_matrix.shape[0]
for zone in zones:
c = int(np.floor(zone / grid_h))
r = int(zone - c * grid_h)
coords_list.append((c, r))
# Solve the TSP optimization problem
tsp_problem = TSPOpt(length=len(coords_list),
coords=coords_list,
maximize=False)
best_path, _ = genetic_alg(tsp_problem, max_iters=10, random_state=2)
best_path = deque(best_path)
# Rotate the path to its best form until
# the starting point P is not 0
P = best_path[0]
while P != 0:
best_path.rotate(1)
P = best_path[0]
return [zones[i] for i in list(best_path)]
def test_random_neighbor():
"""Test random_neighbor method"""
dists = [(0, 1, 3), (0, 2, 5), (0, 3, 1), (0, 4, 7), (1, 3, 6),
(4, 1, 9), (2, 3, 8), (2, 4, 2), (3, 2, 8), (3, 4, 4)]
problem = TSPOpt(5, distances=dists)
x = np.array([0, 1, 2, 3, 4])
problem.set_state(x)
neigh = problem.random_neighbor()
abs_diff = np.abs(x - neigh)
abs_diff[abs_diff > 0] = 1
sum_diff = np.sum(abs_diff)
assert (len(neigh) == 5 and sum_diff == 2 and len(set(neigh)) == 5)
def test_random_mimic():
"""Test random_mimic method"""
dists = [(0, 1, 3), (0, 2, 5), (0, 3, 1), (0, 4, 7), (1, 3, 6),
(4, 1, 9), (2, 3, 8), (2, 4, 2), (3, 2, 8), (3, 4, 4)]
pop = np.array([[1, 0, 3, 2, 4], [0, 2, 1, 3, 4], [0, 2, 4, 3, 1],
[4, 1, 3, 2, 0], [3, 4, 0, 2, 1], [2, 4, 0, 3, 1]])
problem = TSPOpt(5, distances=dists)
problem.keep_sample = pop
problem.eval_node_probs()
problem.find_sample_order()
rand = problem.random_mimic()
assert (len(rand) == 5 and len(set(rand)) == 5)
def generate(seed, number_of_cities, area_width=250, area_height=250):
np.random.seed(seed)
x_coords = np.random.randint(area_width, size=number_of_cities)
y_coords = np.random.randint(area_height, size=number_of_cities)
coords = list(tuple(zip(x_coords, y_coords)))
duplicates = TSPGenerator.list_duplicates_(coords)
while len(duplicates) > 0:
for d in duplicates:
x_coords = np.random.randint(area_width, size=len(d))
y_coords = np.random.randint(area_height, size=len(d))
for i in range(len(d)):
coords[d[i]] = (x_coords[i], y_coords[i])
pass
duplicates = TSPGenerator.list_duplicates_(coords)
distances = TSPGenerator.get_distances(coords, False)
return TSPOpt(coords=coords, distances=distances, maximize=False)
def test_find_neighbors():
"""Test find_neighbors method"""
dists = [(0, 1, 3), (0, 2, 5), (0, 3, 1), (0, 4, 7), (1, 3, 6),
(4, 1, 9), (2, 3, 8), (2, 4, 2), (3, 2, 8), (3, 4, 4)]
problem = TSPOpt(5, distances=dists)
x = np.array([0, 1, 2, 3, 4])
problem.set_state(x)
problem.find_neighbors()
neigh = np.array([[1, 0, 2, 3, 4], [2, 1, 0, 3, 4], [3, 1, 2, 0, 4],
[4, 1, 2, 3, 0], [0, 2, 1, 3, 4], [0, 3, 2, 1, 4],
[0, 4, 2, 3, 1], [0, 1, 3, 2, 4], [0, 1, 4, 3, 2],
[0, 1, 2, 4, 3]])
assert np.array_equal(np.array(problem.neighbors), neigh)
def compute_shortest_path(self, starting_zone_id, other_zone_ids):
zones = [starting_zone_id]
[zones.append(zone) for zone in other_zone_ids]
coords_list = []
for zone in zones:
zone_column = int(np.floor(
zone / self.simInput.grid_matrix.shape[0]
))
zone_row = int(
zone - zone_column * self.simInput.grid_matrix.shape[0]
)
coords_list.append((zone_column, zone_row))
problem = TSPOpt(length=len(coords_list), coords=coords_list, maximize=False)
best_path, _ = genetic_alg(problem, max_iters=10, random_state=2)
best_path = deque(best_path)
starting_point = best_path[0]
while starting_point != 0: # Starting point is not worker position
best_path.rotate(1)
starting_point = best_path[0]
return [zones[i] for i in list(best_path)]
def test_sample_pop():
"""Test sample_pop method"""
dists = [(0, 1, 3), (0, 2, 5), (0, 3, 1), (0, 4, 7), (1, 3, 6),
(4, 1, 9), (2, 3, 8), (2, 4, 2), (3, 2, 8), (3, 4, 4)]
pop = np.array([[1, 0, 3, 2, 4], [0, 2, 1, 3, 4], [0, 2, 4, 3, 1],
[4, 1, 3, 2, 0], [3, 4, 0, 2, 1], [2, 4, 0, 3, 1]])
problem = TSPOpt(5, distances=dists)
problem.keep_sample = pop
problem.eval_node_probs()
sample = problem.sample_pop(100)
row_sums = np.sum(sample, axis=1)
assert (np.shape(sample)[0] == 100 and np.shape(sample)[1] == 5
and max(row_sums) == 10 and min(row_sums) == 10)