64 lines
1.9 KiB
Python
64 lines
1.9 KiB
Python
import numpy as np
|
|
from typing import List
|
|
from matplotlib import pyplot as plt
|
|
from numpy.core._multiarray_umath import ndarray
|
|
|
|
|
|
class Instance:
|
|
nPoints: int
|
|
best_sol: int
|
|
name: str
|
|
lines: List[str]
|
|
dist_matrix: ndarray
|
|
points: ndarray
|
|
|
|
def __init__(self, name_tsp):
|
|
self.read_instance(name_tsp)
|
|
|
|
def read_instance(self, name_tsp):
|
|
# read raw data
|
|
file_object = open(name_tsp)
|
|
data = file_object.read()
|
|
file_object.close()
|
|
self.lines = data.splitlines()
|
|
|
|
# store data set information
|
|
self.name = self.lines[0].split(' ')[2]
|
|
self.nPoints = np.int(self.lines[3].split(' ')[2])
|
|
self.best_sol = np.int(self.lines[5].split(' ')[2])
|
|
|
|
# read all data points and store them
|
|
self.points = np.zeros((self.nPoints, 3))
|
|
for i in range(self.nPoints):
|
|
line_i = self.lines[7 + i].split(' ')
|
|
self.points[i, 0] = line_i[0]
|
|
self.points[i, 1] = line_i[1]
|
|
self.points[i, 2] = line_i[2]
|
|
|
|
self.create_dist_matrix()
|
|
|
|
def print_info(self):
|
|
print('name: ' + self.name)
|
|
print('nPoints: ' + str(self.nPoints))
|
|
print('best_sol: ' + str(self.best_sol))
|
|
|
|
def plot_data(self):
|
|
plt.figure(figsize=(8, 8))
|
|
plt.title(self.name)
|
|
plt.scatter(self.points[:, 1], self.points[:, 2])
|
|
plt.show()
|
|
|
|
@staticmethod
|
|
def distance_euc(zi, zj):
|
|
xi, xj = zi[0], zj[0]
|
|
yi, yj = zi[0], zj[0]
|
|
return round(np.sqrt((xi - xj) ** 2 + (yi - yj) ** 2))
|
|
|
|
def create_dist_matrix(self):
|
|
self.dist_matrix = np.zeros((self.nPoints, self.nPoints))
|
|
|
|
for i in range(self.nPoints):
|
|
for j in range(i, self.nPoints):
|
|
self.dist_matrix[i, j] = self.distance_euc(self.points[i][1:3], self.points[j][1:3])
|
|
self.dist_matrix += self.dist_matrix.T
|
|
|