-
Notifications
You must be signed in to change notification settings - Fork 0
/
evolution.py
159 lines (130 loc) · 5.8 KB
/
evolution.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
import copy
import numpy as np
import pickle
import random
from player import Player
class Evolution:
def __init__(self):
self.game_mode = "Neuroevolution"
# f.close()
def max_min_average_fitness_save(self, players, average):
f = open("result.txt", "a")
result = str(players[0].fitness) + " " + str(average) + " " + str(players[299].fitness) + "\n"
f.write(result)
def next_population_selection(self, players, num_players):
"""
Gets list of previous and current players (μ + λ) and returns num_players number of players based on their
fitness value.
:param players: list of players in the previous generation
:param num_players: number of players that we return
"""
# TODO (Implement top-k algorithm here)
# TODO (Additional: Implement roulette wheel here)
# TODO (Additional: Implement SUS here)
# TODO (Additional: Learning curve)
average = 0
for p in players:
average += p.fitness
average = average/300
# this sorting is for plotting the result
players.sort(key=lambda x: x.fitness, reverse=True)
# this line is used for saving the max and min and average fitness of players
self.max_min_average_fitness_save(players, average)
fitness_list = list()
for p in players:
fitness_list.append(p.fitness ** 2)
players = random.choices(players, weights=fitness_list, k=num_players)
return players[: num_players]
def generate_new_population(self, num_players, prev_players=None):
"""
Gets survivors and returns a list containing num_players number of children.
:param num_players: Length of returning list
:param prev_players: List of survivors
:return: A list of children
"""
first_generation = prev_players is None
if first_generation:
return [Player(self.game_mode) for _ in range(num_players)]
else:
children_list = []
fitness_list = list()
for p in prev_players:
fitness_list.append(p.fitness ** 2)
prev_players = random.choices(prev_players, weights=fitness_list, k=num_players)
half = int(len(prev_players) / 2) - 1
for i in range(half):
# c1 is child number 1 and c2 is child number 2
c1, c2 = self.crossover_operator(prev_players[i * 2], prev_players[i * 2 + 1])
c1 = self.mutation_operator(c1)
c2 = self.mutation_operator(c2)
children_list.append(c1)
children_list.append(c2)
new_generated_players = children_list
return new_generated_players
def clone_player(self, player):
"""
Gets a player as an input and produces a clone of that player.
"""
new_player = Player(self.game_mode)
new_player.nn = copy.deepcopy(player.nn)
new_player.fitness = player.fitness
return new_player
# single point crossover
def crossover_operator(self, p1, p2):
# p1 is the first parent and p2 is the second parent
p1_w1 = p1.nn.w1
p1_w2 = p1.nn.w2
p1_b1 = p1.nn.b1
p1_b2 = p1.nn.b2
p2_w1 = p2.nn.w1
p2_w2 = p2.nn.w2
p2_b1 = p2.nn.b1
p2_b2 = p2.nn.b2
# c1 is child number 1 and c2 is child number 2
c1 = self.clone_player(p1)
c2 = self.clone_player(p2)
counter = 0
for a in p1_w1:
length_a = int(len(a))
half_of_length_a = int(len(a) / 2)
c1.nn.w1[counter] = np.append(a[0:half_of_length_a], p2_w1[counter][half_of_length_a:length_a])
c2.nn.w1[counter] = np.append(p2_w1[counter][0:half_of_length_a], a[half_of_length_a:length_a])
counter = counter + 1
counter = 0
for a in p1_w2:
length_a = int(len(a))
half_of_length_a = int(len(a) / 2)
c1.nn.w2[counter] = np.append(a[0:half_of_length_a], p2_w2[counter][half_of_length_a:length_a])
c2.nn.w2[counter] = np.append(p2_w2[counter][0:half_of_length_a], a[half_of_length_a:length_a])
counter = counter + 1
length_p1_b1 = int(len(p1_b1))
half_of_length_p1_b1 = int(len(p1_b1) / 2)
c1.nn.b1[0:half_of_length_p1_b1] = p1_b1[0:half_of_length_p1_b1]
c1.nn.b1[half_of_length_p1_b1:length_p1_b1] = p2_b1[half_of_length_p1_b1:length_p1_b1]
c2.nn.b1[0:half_of_length_p1_b1] = p2_b1[0:half_of_length_p1_b1]
c2.nn.b1[half_of_length_p1_b1:length_p1_b1] = p1_b1[half_of_length_p1_b1:length_p1_b1]
length_p1_b2 = int(len(p1_b2))
half_of_length_p1_b2 = int((len(p1_b2) / 2))
c1.nn.b2[0:half_of_length_p1_b2] = p1_b2[0:half_of_length_p1_b2]
c1.nn.b2[half_of_length_p1_b2:length_p1_b2] = p2_b2[half_of_length_p1_b2:length_p1_b2]
c2.nn.b2[0:half_of_length_p1_b2] = p2_b2[0:half_of_length_p1_b2]
c2.nn.b2[half_of_length_p1_b2:length_p1_b2] = p1_b2[half_of_length_p1_b2:length_p1_b2]
return c1, c2
def mutation_operator(self, c):
# c is the creature to operate mutation on
c_w1 = c.nn.w1
c_w2 = c.nn.w2
c_b1 = c.nn.b1
c_b2 = c.nn.b2
# nc is the cloned creature(new creature)
nc = self.clone_player(c)
mutate_prob = 0.3
if np.random.uniform(0, 1, 1)[0] < mutate_prob:
nc.nn.w1 = c_w1 + np.random.randn(10, 6)
if np.random.uniform(0, 1, 1)[0] < mutate_prob:
nc.nn.w2 = c_w2 + np.random.randn(1, 10)
if np.random.uniform(0, 1, 1)[0] < mutate_prob:
nc.nn.b1 = c_b1 + np.random.randn(10, 1)
if np.random.uniform(0, 1, 1)[0] < mutate_prob:
nc.nn.b2 = c_b2 + np.random.randn(1, 1)
return nc