generate list of traveling by GA

This commit is contained in:
Mateusz 2020-05-26 23:44:56 +02:00
parent d363120ef4
commit 80b7e32f47
2 changed files with 77 additions and 2 deletions

View File

@ -1,4 +1,38 @@
from random import random, choice
class GeneticAlgorithm:
pass
def __init__(self, firstPopulation, mutationProbability):
self.firstPopulation = firstPopulation
self.mutationProbability = mutationProbability
def selectionModel(self, generation):
max_selected = int(len(generation) / 10)
sorted_by_assess = sorted(generation, key=lambda x: x.fitness)
return sorted_by_assess[:max_selected]
def stopCondition(self, i):
return i == 100
def run(self):
population = self.firstPopulation
population.sort(key=lambda x: x.fitness)
populationLen = len(population)
i = 0
while True:
selected = self.selectionModel(population)
newPopulation = selected.copy()
while len(newPopulation) != populationLen:
child = choice(population).crossover(choice(population))
if random() <= self.mutationProbability:
child.mutation()
newPopulation.append(child)
population = newPopulation
theBestMatch = min(population, key=lambda x: x.fitness)
print("Generation: {} S: {} fitness: {}".format(i, theBestMatch, theBestMatch.fitness))
i += 1
if self.stopCondition(i):
return theBestMatch

View File

@ -1,10 +1,51 @@
from random import randint, sample
from math import sqrt
from src.AI.GaTravelingForHerbs.GeneticAlgorithm import GeneticAlgorithm
START_COORD = [(6, 2)]
END_COORD = [(10, 7)]
COORDS = [(12, 2), (16, 2), (17, 5), (14, 7), (17, 17), (13, 17), (5, 15), (2, 9), (8, 5), (11, 10)]
class Traveling:
def __init__(self, coords):
self.coords = coords
self.coords = coords
self.fitness = self.evaluate()
def evaluate(self):
sum = 0
for i, c in enumerate(self.coords):
if i + 1 > len(self.coords) - 1:
break
nextCoord = self.coords[i + 1]
# calculate distance
sum += sqrt((nextCoord[0] - c[0]) ** 2 + (nextCoord[1] - c[1]) ** 2)
return sum
def crossover(self, element2: 'Element') -> 'Element':
childCoords = self.coords[1:int(len(self.coords) / 2)]
for coord in element2.coords:
if coord not in childCoords and coord not in END_COORD + START_COORD:
childCoords.append(coord)
if len(childCoords) == len(element2.coords):
break
return Traveling(START_COORD + childCoords + END_COORD)
def mutation(self):
first = randint(1, len(self.coords) - 2)
second = randint(1, len(self.coords) - 2)
self.coords[first], self.coords[second] = self.coords[second], self.coords[first]
self.fitness = self.evaluate()
def __repr__(self):
return str(self.coords)
firstGeneration = [Traveling(START_COORD + sample(COORDS, len(COORDS)) + END_COORD) for _ in range(100)]
mutationProbability = float(0.1)
ga = GeneticAlgorithm(firstGeneration, mutationProbability)
movementList = ga.run()