112 lines
4.1 KiB
Python
112 lines
4.1 KiB
Python
|
import random
|
||
|
|
def geneticAligouwu(whereMines):
|
||
|
|
populationSize = 120
|
||
|
|
topSize = 20
|
||
|
|
mutationCount = 5
|
||
|
|
genCount = 500
|
||
|
|
|
||
|
|
def distance(xfrom, yto):
|
||
|
|
#print(xfrom[0],yto[0],xfrom[1],yto[1])
|
||
|
|
return abs(xfrom[0]-yto[0]) + abs(xfrom[1]-yto[1])
|
||
|
|
#distance ^
|
||
|
|
|
||
|
|
def fitness(route):
|
||
|
|
routescore = 0
|
||
|
|
startingPoint = (0,0)
|
||
|
|
routescore += distance(startingPoint, route[0])
|
||
|
|
for i in range(0, len(route)-1):
|
||
|
|
if i != (len(route)-2):
|
||
|
|
routescore += distance(route[i], route[i+1])
|
||
|
|
return routescore
|
||
|
|
#fitness ^
|
||
|
|
|
||
|
|
def evaluation(population):
|
||
|
|
for i in range(0, len(population)):
|
||
|
|
#print(str(population[i]) + " nr: " + str(i))
|
||
|
|
population[i][-1] = fitness(population[i])
|
||
|
|
#evaluation(distance)
|
||
|
|
|
||
|
|
def Davis(p1, p2):
|
||
|
|
child = []
|
||
|
|
childP1 = []
|
||
|
|
childP2 = []
|
||
|
|
|
||
|
|
geneA = int(random.random() * len(p1))
|
||
|
|
geneB = int(random.random() * len(p1))
|
||
|
|
|
||
|
|
start = min(geneA, geneB)
|
||
|
|
end = max(geneA, geneB)
|
||
|
|
|
||
|
|
for i in range(start, end):
|
||
|
|
childP1.append(p1[i])
|
||
|
|
|
||
|
|
childP2 = [item for item in p2 if item not in childP1]
|
||
|
|
|
||
|
|
child = childP1 + childP2
|
||
|
|
return child
|
||
|
|
#breeding
|
||
|
|
|
||
|
|
|
||
|
|
def produceNextGeneration(population):
|
||
|
|
#print(population[0])
|
||
|
|
evaluation(population)
|
||
|
|
population.sort(key=lambda eme: eme[-1])
|
||
|
|
"""tempprintscore = []
|
||
|
|
for i in range(0, 10):
|
||
|
|
tempprintscore.append(population[i][-1])
|
||
|
|
print(tempprintscore)"""
|
||
|
|
#print("Best dist: " + str(population[0][-1]))
|
||
|
|
#sorting sorted
|
||
|
|
|
||
|
|
newGeneration = []
|
||
|
|
for i in range(0,topSize):
|
||
|
|
newGeneration.append(population[i])
|
||
|
|
#print("elite child: " + str(population[i]) + str(i))
|
||
|
|
#elitism
|
||
|
|
for i in range(0, len(population)-topSize):
|
||
|
|
child = Davis(population[i], population[len(population)-i-1])
|
||
|
|
newGeneration.append(child)
|
||
|
|
#print("breed child: " + str(child) + "nr: " + str(i+topSize) + " / " + str(i))
|
||
|
|
#fill rest of generation with breeding
|
||
|
|
for i in range(mutationCount):
|
||
|
|
mO = int(random.random() * len(population))
|
||
|
|
#MutatingOrganism
|
||
|
|
gene1 = int(abs(random.random() * len(newGeneration[mO])-1))
|
||
|
|
gene2 = int(abs(random.random() * len(newGeneration[mO])-1))
|
||
|
|
geneTemp = newGeneration[mO][gene1]
|
||
|
|
newGeneration[mO][gene1] = newGeneration[mO][gene2]
|
||
|
|
newGeneration[mO][gene2] = geneTemp
|
||
|
|
#mutation by swapping mines
|
||
|
|
return newGeneration
|
||
|
|
|
||
|
|
|
||
|
|
def geneticAlgorithm(population, generations):
|
||
|
|
currentBestRoute = []
|
||
|
|
currentHighScore = 9999999999999
|
||
|
|
for i in range(0, generations):
|
||
|
|
population = produceNextGeneration(population)
|
||
|
|
if population[0][-1] < currentHighScore:
|
||
|
|
currentBestRoute = population[0]
|
||
|
|
|
||
|
|
print("Najlepszy dystans: " + str(population[0][-1]) + " / " + str(currentBestRoute[-1]) + " <3")
|
||
|
|
return currentBestRoute
|
||
|
|
|
||
|
|
|
||
|
|
#whereMines = []
|
||
|
|
#for encounter in self.current_map.encounters:
|
||
|
|
# whereMines.append((encounter.position_x, encounter.position_y))
|
||
|
|
#mines(coordinates) ^
|
||
|
|
population = []
|
||
|
|
for i in range(0, populationSize):
|
||
|
|
tempPopulation = random.sample(whereMines, len(whereMines))
|
||
|
|
tempPopulation.append(0)
|
||
|
|
population.append(tempPopulation)
|
||
|
|
#generation
|
||
|
|
####################################################################################
|
||
|
|
|
||
|
|
|
||
|
|
route = geneticAlgorithm(population, genCount)
|
||
|
|
route.pop()
|
||
|
|
#END OF GENETIC ALGORITHM
|
||
|
|
####################################################################################
|
||
|
|
return route
|