Genetics

genetic.py

@@ -0,0 +1,123 @@
+import random
+
+
+class genetic:
+    def __init__(self, chrome):
+        self.chrome = chrome
+        self.key = 10
+        self.pop_size = 50
+        self.gen_max = 20
+        self.length = len(chrome) - 1
+        self.div = self.length // 2
+        self.rand = list(range(1, self.length + 1))
+
+    def create_genome(self):
+        genome = "0"
+        exclusive = random.sample(self.rand, self.length)
+        genome += "".join(map(str, exclusive))
+        return genome
+
+    def find_fitness(self, genome):
+        fitness = 0
+        for i in range(len(genome) - 1):
+            fitness += self.chrome[int(genome[i])][int(genome[i + 1])]
+        return fitness
+
+    def crossOver(self, gen1, gen2):
+        gen = []
+        for i in range(self.div):
+            gen.append(gen1[i])
+        for i in range(self.div, self.length+1):
+            if gen2[i] not in gen:
+                gen.append(gen2[i])
+            else:
+                j = 0
+                while gen1[j] in gen:
+                    j += 1
+                gen.append(gen1[j])
+        return ''.join(gen)
+
+    def mutation(self, genome):
+        genome = list(genome)
+        while True:
+            gen1 = random.randint(1, self.length - 1)
+            gen2 = random.randint(1, self.length - 1)
+            if gen1 != gen2:
+                genome[gen1], genome[gen2] = genome[gen2], genome[gen1]
+                break
+        return ''.join(genome)
+
+    def search(self):
+        gen = 1
+        population = []
+
+        for i in range(self.pop_size):
+            ind = individual()
+            ind.genome = self.create_genome()
+            ind.fitness = self.find_fitness(ind.genome)
+            population.append(ind)
+
+        print("First population:")
+        for ind in population:
+            print(ind.genome, ind.fitness)
+
+        while gen <= self.gen_max:
+            new_population = population
+
+            key = random.randint(0, 100)
+            if key > self.key:
+                for i in range(self.pop_size):
+                    temp1, temp2 = random.sample(population, 2)
+                    new_ind1 = individual()
+                    new_ind1.genome = self.crossOver(temp1.genome, temp2.genome)
+                    new_ind1.fitness = self.find_fitness(new_ind1.genome)
+                    new_ind2 = individual()
+                    new_ind2.genome = self.crossOver(temp2.genome, temp1.genome)
+                    new_ind2.fitness = self.find_fitness(new_ind2.genome)
+                    new_population.append(new_ind1)
+                    new_population.append(new_ind2)
+                    new_population.sort()
+                    new_population.pop()
+                    new_population.pop()
+            else:
+                print('mutation')
+                for i in range(self.pop_size):
+                    temp1 = random.choice(population)
+                    new_ind = individual()
+                    new_ind.genome = self.mutation(temp1.genome)
+                    new_ind.fitness = self.find_fitness(new_ind.genome)
+                    new_population.append(new_ind)
+                    new_population.sort()
+                    new_population.pop()
+            print("\n")
+            print("New Generation # ", gen)
+            print("Genome	 Fitness")
+
+            for i in range(self.pop_size):
+                print(new_population[i].genome, new_population[i].fitness)
+
+            print("\n")
+
+            gen += 1
+
+        ans = min(population, key=lambda x: x.fitness)
+        print("Final order: ", ans.genome, ans.fitness)
+        return ans.genome
+
+
+class individual:
+    def __init__(self):
+        self.genome = 0
+        self.fitness = 0
+
+    def __lt__(self, other):
+        return self.fitness < other.fitness
+
+    def __gt__(self, other):
+        return self.fitness > other.fitness
+
+    def __le__(self, other):
+        return self.fitness <= other.fitness
+
+    def __ge__(self, other):
+        return self.fitness >= other.fitness

main.py

@@ -7,6 +7,7 @@ from rubbish import *
 from tree import evaluate_values, trash_selection
 from truck import Truck
 from surface import *
+from genetic import genetic
 
 RESOLUTION = 900
 SIZE = 60
@@ -52,6 +53,11 @@ for i in range(15):
             rubbish_list.append(Rubbish(screen, j * 60, i * 60))
 
 path = []
+gen = [(truck.y / 60, truck.x / 60)]
+fl = 0
+length = []
+finalLength = []
+order = []
 while True:
     pygame.time.delay(500)
 
@@ -64,11 +70,40 @@ while True:
         i.draw_rubbish()
     truck.draw_truck()
 
+    # finding order to collect rubbish
+    if fl == 0:
+        for item in rubbish_list:
+            print(item.y / 60, item.x / 60, end='\n')
+            gen.append((item.y / 60, item.x / 60))
+        for item1 in range(len(gen)):
+            for item2 in range(len(gen)):
+                if item1 < item2:
+                    length.append(len(a_star(surface_list, gen[item2]).tree_search(PriorityQueue(), gen[item1], 'R')))
+                else:
+                    length.append(0)
+            finalLength.append(length)
+            length = []
+        fl = 1
+        for i in range(len(finalLength)):
+            for j in range(len(finalLength)):
+                if i > j:
+                    finalLength[i][j] = finalLength[j][i]
+        for i in range(len(finalLength)):
+            for j in range(len(finalLength)):
+                print(finalLength[i][j], end=',')
+            print('')
+        print(finalLength)
+        order = genetic(finalLength).search()
+        order = list(map(int, order))
+        order.pop(0)
+        for j in range(len(order)):
+            order[j] -= 1
+
     # finding a path to rubbish
-    if rubbish_list and not path:
+    if order and not path:
         start = (truck.y / 60, truck.x / 60)
         direction = truck.direction
-        currentRubbish = rubbish_list[0]
+        currentRubbish = rubbish_list[order[0]]
         endpoint = (currentRubbish.y / 60, currentRubbish.x / 60)
         # path = bfs(surface_list, endpoint).tree_search(deque(), start, direction)
         path = a_star(surface_list, endpoint).tree_search(PriorityQueue(), start, direction)
@@ -82,8 +117,8 @@ while True:
             truck.change_direction(action)
 
     # the decision that takes what to do with the garbage
-    if not path and rubbish_list:
-        data = rubbish_list[0].data_for_decision_tree()
+    if not path and order:
+        data = rubbish_list[order[0]].data_for_decision_tree()
         print(f'----------\n'
               f'Characteristics of the garbage we met:\n'
               f'Weight:{data[0]}\nDensity:{data[1]}\n'
@@ -94,12 +129,11 @@ while True:
         decision = trash_selection(evaluate_values(data))
         if decision == [0]:
             print('We refused this rubbish because of bad characteristics')
-            rubbish_list[0].rubbish_refused()
-            refused_rubbish_list.append(rubbish_list[0])
+            rubbish_list[order[0]].rubbish_refused()
+            refused_rubbish_list.append(rubbish_list[order[0]])
         else:
             print('We take this rubbish because of good characteristics')
-        rubbish_list.pop(0)
-
+        order.pop(0)
     pygame.display.flip()
 
     for event in pygame.event.get():