def do_algorithm(self, distance_matrix):
vertex_list = np.array([np.random.randint(len(distance_matrix))])
shortest_cycle = Cycle(vertex_list, Cycle.calculate_cycle_length(distance_matrix,
vertex_list))
while len(vertex_list) / len(distance_matrix) < GreedyCycleStrategy.VERTEX_RATIO:
shortest_cycle = GreedyCycleStrategy.__find_shortest_cycle(distance_matrix, vertex_list)
vertex_list = shortest_cycle.cycle
return shortest_cycle
def __find_nearest_neighbor(distance_matrix, vertex_list):
remaining_vertex_list = np.delete(np.arange(len(distance_matrix)), vertex_list)
last_vertex = vertex_list[-1]
distance_list = np.copy(distance_matrix[last_vertex])
remaining_distance_list = np.copy(distance_list[remaining_vertex_list])
min_length = np.min(remaining_distance_list)
nearest_neighbor = np.where(distance_list == min_length)[0][0]
new_vertex_list = np.append(vertex_list, nearest_neighbor)
return Cycle(new_vertex_list, Cycle.calculate_cycle_length(distance_matrix,
new_vertex_list))
def __find_shortest_cycle(distance_matrix, vertex_list):
remaining_vertex_list = np.delete(np.arange(len(distance_matrix)), vertex_list)
cycle_list = np.array([])
for vertex in remaining_vertex_list:
new_vertex_list = np.append(vertex_list, vertex)
cycle_list = np.append(cycle_list,
Cycle(new_vertex_list,
Cycle.calculate_cycle_length(distance_matrix,
new_vertex_list)))
return min(cycle_list, key=lambda cycle: cycle.length)
def main():
distance_matrix = np.array([[0, 1, 2, 3, 4], [1, 0, 2, 3, 4],
[2, 2, 0, 3, 4], [3, 3, 3, 0, 4],
[4, 4, 4, 4, 0]])
vertex_list = np.array([0, 1, 2])
length = Cycle.calculate_cycle_length(distance_matrix, vertex_list)
strategy = RegretStrategy()
result = strategy.do_algorithm(distance_matrix)
print(length)
print(result)
def __find_max_regret_cycle(distance_matrix, vertex_list):
regrets_list = np.array([])
remaining_vertex_list = np.delete(np.arange(len(distance_matrix)),
vertex_list)
for vertex in remaining_vertex_list:
regret = RegretStrategy.__calculate_regret(
distance_matrix, vertex,
np.delete(remaining_vertex_list,
np.where(remaining_vertex_list == vertex)))
regrets_list = np.append(regrets_list, {
'vertex': vertex,
'regret': regret
})
max_regret = max(regrets_list,
key=lambda regret_dict: regret_dict['regret'])
new_vertex_list = np.append(vertex_list, max_regret['vertex'])
return Cycle(
new_vertex_list,
Cycle.calculate_cycle_length(distance_matrix, new_vertex_list))
def do_algorithm(self, distance_matrix):
vert_list = np.random.randint(0, 100, 10)
vert_list = np.append(vert_list, vert_list[0])
return Cycle(vert_list, random.randint(0, 1000))