def calcFitnesses(self):
'''
calculate fitness of every chromosome
'''
costs = []
routes = []
for c in range(self.n_chromosomes):
cost = 0
truck_routes = []
for t in range(self.n_trucks):
if (len(self.truck_loads[c][t]) > 0):
if (len(self.truck_loads[c][t]) <= 9):
truck_distance, path = self.brute_force_solution(
self.truck_loads[c][t])
else:
truck_distance_mat = self.submat(
self.truck_loads[c][t])
ac = AntColony(distances=truck_distance_mat,
n_ants=20,
n_best=10,
n_iterations=100,
decay=0.95,
alpha=1,
beta=4)
path, truck_distance = ac.run()
cost += (truck_distance * self.transportation_costs[t])
truck_routes.append(path)
else:
truck_routes.append(())
costs.append(cost)
routes.append(truck_routes)
self.costs = np.array(costs)
self.fitnesses = (self.fitness_counter - self.costs) + 1
self.routes = routes
def calcFitnessesOffspring(self):
'''
calculate fitness of every chromosome
'''
costs = []
routes = []
for c in range(np.shape(self.offspring)[2]):
cost = 0
truck_routes = []
for t in range(self.n_trucks):
if (len(self.offspring_loads[c][t]) > 0):
truck_distance_mat = self.submat(
self.offspring_loads[c][t])
if (len(self.offspring_loads[c][t]) > 3):
ac = AntColony(distances=truck_distance_mat,
n_ants=20,
n_best=10,
n_iterations=10,
decay=0.95,
alpha=1,
beta=4)
path, truck_distance = ac.run()
else:
truck_distance = sum([
truck_distance_mat[i, (i + 1) %
len(truck_distance_mat)]
for i in range(len(truck_distance_mat))
])
path = [(i, (i + 1) % len(truck_distance_mat))
for i in range(len(truck_distance_mat))]
cost += (truck_distance * self.transportation_costs[t])
truck_routes.append(path)
else:
truck_routes.append(())
costs.append(cost)
routes.append(truck_routes)
self.offspring_costs = np.array(costs)
self.offspring_fitnesses = (self.fitness_counter - self.costs) + 1
self.routes = routes
import numpy as np
from ant_colony import AntColony
distances = np.array([[np.inf, 2, 2, 5, 7], [2, np.inf, 4, 8, 2],
[2, 4, np.inf, 1, 3], [5, 8, 1, np.inf, 2],
[7, 2, 3, 2, np.inf]])
ant_colony = AntColony(distances, 1, 1, 100, 0.95, alpha=1, beta=1)
shortest_path = ant_colony.run()
print("shorted_path: {}".format(shortest_path))
Ant Colony Optimization stolen from: https://raw.githubusercontent.com/Akavall/AntColonyOptimization/master/ant_colony.py ''' n_ants = 100 n_best = 25 n_iterations = 300 decay = 0.1 distCopy = distances.copy() np.fill_diagonal(distCopy,np.inf) colony = AntColony(distCopy, n_ants, n_best, n_iterations, decay, alpha=1.0, beta=1.5,coords=coords) shortes_path = colony.run() # only city indices route = [s[1]-1 for s in shortes_path[0][:-1]] # demands sorted by traveling route demandRoute = demands[route] # genetic algorithm gA = GenAlgo(coords, demands, route, demandRoute, distances[1:,1:], depotDistances, vehicleCapacities, vehicleCosts) routes = [] demandRoutes = [] topCosts = [] assignments = []
# Made by - Pratik Luitel , Anish Dulal, Stuti Kafle
# available at - https://github.com/pratikluitel/ACOVisualization
# make a venv and install dependencies in requirements.txt before you proceed
import numpy as np
from ant_colony import AntColony
# making a random large city with 100 streets
large_distances = np.random.randint(1, 10, size=(100, 100))
large_distances = (large_distances + large_distances.T).astype('float64')
np.fill_diagonal(large_distances, float('inf'))
large_ant_colony = AntColony(large_distances, 10, 5, 50, 0.15, alpha=1, beta=1)
print(large_ant_colony.run())
# Made by - Pratik Luitel , Anish Dulal, Stuti Kafles
# available at - https://github.com/pratikluitel/ACOVisualization
# make a venv and install dependencies in requirements.txt before you proceed
import numpy as np
from ant_colony import AntColony
# making a city with 5 streets
small_distances = np.array([[np.inf, 2, 2, 5, 7, 4], [2, np.inf, 4, 8, 2, 3],
[2, 4, np.inf, 1, 3, 6], [5, 8, 1, np.inf, 2, 8],
[7, 2, 3, 2, np.inf, 1], [4, 3, 6, 8, 1, np.inf]])
small_ant_colony = AntColony(small_distances,
5,
2,
10,
0.95,
alpha=1,
beta=1,
graph=False)
small_ant_colony.run()