plt.plot(progress)
plt.ylabel('Distance')
plt.xlabel('Generation')
plt.subplot(212)
plt.plot(progress_min)
plt.ylabel('Minimum Distance')
plt.xlabel('Generation')
plt.show()
print("Minimum distance :", min_dist)
print("Best route :", bestRoute)
#INPUT
import sys
if len(sys.argv) < 2:
print("need inpute file")
sys.exit(1)
r = ReadData(sys.argv[1])
print(r.size)
size = r.size
dist_matrix = r.GetDistanceMat()
cityList = []
for i in range(size):
cityList.append(City(label=i + 1, distance_list=dist_matrix[i]))
geneticAlgorithmPlot(population=cityList,
popSize=500,
eliteSize=100,
mutationRate=0.01,
generations=100)
class NN():
name = "Nearest Neighbor"
def __init__(self, file):
"""
Intialize: Instaces file,
Distance Matrix,
and size
"""
self.file = file
self.instance = ReadData(self.file)
self.size = self.instance.size
self.dis_mat = self.instance.GetDistanceMat()
self.time_read = self.instance.time_to_read
self.time_algo = 0
def get_dist_mat(self):
"""
changing the distances between tha
i to i with infinity for simility in code
"""
D = self.dis_mat.copy()
for i in range(self.size):
D[i][i] = np.inf
return D
def nn_algo(self, startPoint):
"""
Nearest Neighbour algorithm
"""
dist_mat = self.get_dist_mat()
Tour = [startPoint]
for _ in range(self.size - 1):
min_index = np.argmin(dist_mat[Tour[-1]])
for t in Tour:
dist_mat[min_index][t] = np.inf
dist_mat[t][min_index] = np.inf
Tour.append(min_index)
return np.array(Tour)
def run(self):
"""
randomly chooces starting point
10% of intances size
and max = 1000
min = 10
and call nn_algo for tour
"""
tours_dist = []
tours = []
self._write_info()
startPoints = self._start_pt_list()
#print(startPoints)
#startPoints = 0
for s in startPoints:
t = self.nn_algo(s)
d = self.get_tour_distance(t)
tours.append(t + 1)
tours_dist.append(d)
self._best_tour(tours, tours_dist)
def _best_tour(self, Ts, Tsd):
min_dist_index = np.argmin(Tsd)
self._writestat(Tsd[min_dist_index], Ts[min_dist_index])
def get_tour_distance(self, T):
s = 0
for i, t in enumerate(T):
try:
s += self.dis_mat[t][T[i + 1]]
except IndexError:
s += self.dis_mat[t][T[0]]
return s
def _write_info(self):
"""
write info about instance
"""
print("Instance name:", self.instance.name)
print("Dimention:", self.size)
print("Distance Type:", self.instance.EdgeWeightType)
print("\n \t \t Running 2-opt over 50 random tour ")
def _writestat(self, D, T):
"""
Write stats
"""
print("\n Tour Distance: ", D)
print(" Best Tour by 2-opt is: \n", T)
print("\n Time to read instance (sec): ", round(self.time_read))
self.time_algo = time.time() - start_time
print(" Time to run instances(sec): ", round(self.time_algo))
print(" Total Time (sec): ", round(self.time_read + self.time_algo))
def _start_pt_list(self):
"""
return starting points list
"""
np.random.seed(1)
a = round(self.size * 0.1)
mi = 10
mx = 1000
if a > mx:
l = np.random.choice(self.size, mx, replace=False)
return (l)
elif a <= 10:
l = np.random.choice(self.size, mi, replace=False)
return (l)
else:
l = np.random.choice(self.size, a, replace=False)
return (l)
class TwoOPT:
"""
2-opt:
Generate intial tour
and improve it by deleting 1 one edges
and change with other
-- It gives nearly optimal tour
"""
def __init__(self, file):
"""
Intialize: Instaces file,
Distance Matrix,
and size
"""
self.file = file
self.instance = ReadData(self.file)
self.size = self.instance.size
self.dis_mat = self.instance.GetDistanceMat()
self.time_read = self.instance.time_to_read
self.time_algo = 0
def get_initial_tour(self):
"""
Return: intial tour
"""
return [*range(1, self.size + 1)]
def Swap(self, tour, x, y):
"""
tour : Given TSP tour
x = swappping First index in tour
y = swappping last index in tour
return : new_tour with perfomming swapping
note: x and y should be index only (in tour) not exact city number
"""
new_tour = tour[:x] + [*reversed(tour[x:y + 1])] + tour[y + 1:]
return new_tour
def get_distance(self, tour):
"""
Given any tour it return total distance of
given tour
dis_mat : distance matrix
"""
total_dis = 0
for ind, r in enumerate(tour):
_from = r
if ind + 1 == len(tour):
_to = tour[0]
total_dis += self.dis_mat[_from - 1][_to - 1]
else:
_to = tour[ind + 1]
total_dis += self.dis_mat[_from - 1][_to - 1]
return total_dis
def _optimize(self, initial_tour, Debuglevel=0):
"""
Improve existing tour
using 2-opt method
"""
minchange = -1
tour = initial_tour
while minchange < 0:
minchange = 0
for i in range(self.size - 3):
for j in range(i + 2, self.size - 1):
t1 = tour[i]
t2 = tour[i + 1]
t3 = tour[j]
t4 = tour[j + 1]
change = (self.dis_mat[t1 - 1][t3 - 1] +
self.dis_mat[t2 - 1][t4 - 1] -
self.dis_mat[t1 - 1][t2 - 1] -
self.dis_mat[t3 - 1][t4 - 1])
if change < minchange:
minchange = change
tour = self.Swap(tour, i + 1, j)
if Debuglevel:
print("Tour After Change : ", minchange, "Distances: ",
self.get_distance(tour))
self.best_tour = tour
return tour
def _initial_random_tour(self,seed):
""""
Return randomly generated tour
"""
np.random.seed(seed)
T = np.arange(1,self.size+1)
np.random.shuffle(T)
return list(T)
def run(self):
tours = []
tours_dist = []
#self._write_info()
for r in range(1):
T = self._initial_random_tour(r)
tour = self._optimize(T)
tour_distance = self.get_distance(tour)
tours.append(tour)
tours_dist.append(tour_distance)
min_dist_index = np.argmin(tours_dist)
return (tours_dist[min_dist_index], tours[min_dist_index])
"""def _write_info(self):
from readData import ReadData
from SA2opt import SimAnneal
import matplotlib.pyplot as plt
import random
import numpy as np
import sys
if len(sys.argv)<2:
print("need inpute file")
sys.exit(1)
filename = sys.argv[1]
D = ReadData(filename)
if __name__ == "__main__":
# coords = [[random.uniform(-1000, 1000), random.uniform(-1000, 1000)] for i in range(100)]
sa = SimAnneal(D.GetDistanceMat(),filename)#, stopping_iter=2)
sa.anneal()
#sa.batch_anneal()
#sa.visualize_routes()
sa.plot_learning()
