/
GA.py
222 lines (193 loc) · 10.5 KB
/
GA.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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
#!/usr/bin/env python
import sys
import os
import time
import random
import pprocess
import individual
class RunGA():
def __init__(self, rep_length, popsize, sp, mut, fitfun, maxgen, cross, nr_processes, run_id , path, onlyThebest, runval):
self.nr_processes = nr_processes
self.currentIteration=0
self.fitfun =fitfun
self.maxIterations=maxgen
self.popsize=popsize
self.cars = []
self.fitArray = []
self.onlyTheBest = onlyThebest
if self.onlyTheBest == 0:
for i in range(self.popsize):
self.cars.append(individual.individual(random.uniform(50,150), random.uniform(0,0.1), random.uniform(0,20), random.uniform(0,1), random.uniform(0,100), random.uniform(0,1), random.uniform(0,1), random.uniform(0,70), random.uniform(0,1), random.randint(5000,8000),random.uniform(0,1),random.uniform(50,110),random.uniform(50,110),random.uniform(5,15)))
print len(self.cars)
if self.onlyTheBest == 1:
#total:1339.172; aalborg: 100.808 alpine1: 212.258 alpine2: 143.42 brondehach: 135.83 corkscrew: 120.926 eroad: 95.704 gtrack: 61.342 forza: 156.748 wheel1: 128.67 wheel2: 183.466 parameters: ['154.897575174', '-0.0199', '-0.139884752072', '0.779904990626', '82.8097466274', '0.971351422208', '0.246257782509', '32.9286263315', '0.0215211462875', '5909.19902', '0.829675710924', '78.4693515643', '49.7170137868', '17.9086116398']
#FAEHRT SICHER aber auf manchen strecken ein wenig langsammer
self.cars.append(individual.individual('154.897575174', '-0.0199', '-0.139884752072', '0.779904990626', '82.8097466274', '0.971351422208', '0.246257782509', '32.9286263315', '0.0215211462875', '5909.19902', '0.829675710924', '78.4693515643', '49.7170137868', '17.9086116398'))
#total:1314.005; aalborg: 100.144 alpine1: 202.608 alpine2: 139.396 brondehach: 131.636 corkscrew: 123.965 eroad: 95.432 gtrack: 61.22 forza: 154.07 wheel1: 128.66 wheel2: 176.874 parameters: ['154.897575174', '-0.0199', '-0.0538732272792', '0.779904990626', '82.8097466274', '0.774216279987', '0.246257782509', '32.9286263315', '0.0215211462875', '5909.19902', '0.829675710924', '78.4693515643', '44.6453124081', '17.9086116398']
#SCHNELLER aber unsicherer
#self.cars.append(individual.individual('154.897575174', '-0.0199', '-0.0538732272792', '0.779904990626', '82.8097466274', '0.774216279987', '0.246257782509', '32.9286263315', '0.0215211462875', '5909.19902', '0.829675710924', '78.4693515643', '44.6453124081', '17.9086116398'))
self.tournamentSize=sp
self.crossoverChance = cross
self.debug=0
self.mutationChance = mut
self.runval=runval
def step(self):
pass
def stop(self):
pass
def run(self):
while True:
self.step()
if self.stop():
break
return
def get_current_iteration(self):
return self.current_iteration
class gax(RunGA):
def __init__(self, *args, **kwargs):
RunGA.__init__(self, *args,**kwargs)
self.stop_reached = False
def step(self):
self.tempArray = [i for i in range(len(self.cars))]
self.getFitnessValues()
for replacer in range (len(self.cars)):
self.numberOfElites = (int((1-self.crossoverChance)*self.popsize))+2
if replacer < self.numberOfElites: ##keep the elites
self.tempArray[replacer] = self.cars[self.returnFittest()]
pass
else:
for i in range(0,self.tournamentSize):##select parents in tournament
self.parent1 = 9999999
self.parent2 = 9999999
self.randomPos = random.randint(0, len(self.fitArray)-1)
while(self.fitArray[self.randomPos] == -1):
#print "-1.. get a new one"
self.randomPos = random.randint(0, len(self.fitArray)-1)
if self.parent1==9999999:
self.parent1=self.randomPos
elif self.fitArray[self.randomPos] < self.fitArray[self.parent1]:
if self.fitArray[self.randomPos] == -1:
pass
else:
self.parent1=self.randomPos
self.randomPos = random.randint(0, len(self.fitArray)-1)
while(self.fitArray[self.randomPos] == -1):
self.randomPos = random.randint(0, len(self.fitArray)-1)
if self.parent2==9999999:
self.parent2=self.randomPos
elif self.fitArray[self.randomPos] < self.fitArray[self.parent2]:
if self.fitArray[self.randomPos] == -1:
pass
else:
self.parent2=self.randomPos
self.tempArray[replacer]=individual.individual(self.cars[self.parent1].values[0],self.cars[self.parent2].values[1], self.cars[self.parent1].values[2], self.cars[self.parent2].values[3], self.cars[self.parent1].values[4], self.cars[self.parent2].values[5], self.cars[self.parent1].values[6], self.cars[self.parent2].values[7],self.cars[self.parent1].values[8],self.cars[self.parent2].values[9],self.cars[self.parent1].values[10],self.cars[self.parent2].values[11],self.cars[self.parent1].values[12], self.cars[self.parent2].values[13] )
for i in range(len(self.cars)):
self.cars[i] = self.tempArray[i]
for i in range(self.numberOfElites, len(self.cars)):
for gene in range(14):
# print "mutating "+str(i)
if (random.uniform(0, 1) <= self.mutationChance):
muval = (self.cars[i].values[gene] * 0.1)+0.1 #random.uniform(-0.1,0.1)
if (random.uniform(0,1) <= 0.5):
muval = muval*-1
self.cars[i].values[gene] = self.cars[i].values[gene] + muval
self.cars[i].parameters[gene] = str(float(self.cars[i].parameters[gene]) + muval)
pass
self.cars[0]=self.cars[self.returnFittest()]
if self.currentIteration == 0:
print "----------------------------------------------"
print "run "+str(self.runval)+" starts here"
print "run: "+str(self.runval)+" generation: "+str(self.currentIteration)+" fittest at: "+str(self.returnFittest())+" its fitness: "+str(self.returnFittestFitness())
print "it's values: "+str(self.returnFittestValues())
print "pop avg fitn: "+str(self.returnAvgFitness())
print "fitnessarray: "+str(self.fitArray)
print "number of invalids: "+str(self.invalid)
print"\n\n"
bashCommand = "killall torcs-bin"
os.system(bashCommand)
print "killed old torcs processes"
f = open("debugline_wheel2_aal_forza_eroad_street_alpine1.csv","a")
f.write("run: "+str(self.runval)+" generation: "+str(self.currentIteration)+" fittest at: "+str(self.returnFittest())+" its fitness: "+str(self.returnFittestFitness())+"\n")
f.close
f = open("june14_aal_wheel2_aal_forza_eroad_street_alpine1.csv","a") #opens file with name of "test.txt"
if self.currentIteration == 0:
f.write(("\n\n\n\n"))
f.write(("run "+str(self.runval)+" starts here\n"))
f.write("generation; "+str(self.currentIteration)+"; ")
f.write("it's values; "+str(self.returnFittestValues())+"; number of invalids;"+str(self.invalid)+" ; ")
f.write("fitnessarray; "+str(self.fitArray)+"; ")
f.write("pop avg fitn; "+str(self.returnAvgFitness())+"; ")
f.write("fittest at; "+str(self.returnFittest())+"; its fitness; "+str(self.returnFittestFitness())+";\n ")
if self.currentIteration >= self.maxIterations:
f.write(("the end:\n\n\n\n\n"))
f.close
self.currentIteration=self.currentIteration+1
if self.currentIteration >= self.maxIterations:
self.stop_reached = True
def stop(self):
return self.stop_reached
pass
def printPop(self):
for i in range (len(self.cars)):
print self.cars[i].getParameters()
pass
def getFitnessValues(self):
#print "making fitness"
self.fitArray = [i for i in range(len(self.cars))]
if self.debug == 1:
for i in range(len(self.fitArray)):
self.fitArray[i]= random.uniform(50,500)
else:
self.fitArray = self.evaluate()
def printFitnessArray(self):
if self.fitArray == []:
print "No fitness values collected, yet"
else:
for i in range(len(self.fitArray)):
print self.fitArray[i]
def returnFittest(self):
leader=1000
for i in range(len(self.fitArray)):
#print "i am here"
# print self.fitArray[i]
if leader == 1000:
if self.fitArray[i] != -1:
leader = i
elif (self.fitArray[i] < self.fitArray[leader]) and leader!=1000:
if self.fitArray[i] != -1:
leader = i
return leader
def returnFittestValues(self):
return self.cars[self.returnFittest()].getParameters()
def returnFittestFitness(self):
return self.fitArray[self.returnFittest()]
def returnAvgFitness(self):
total = 0
self.invalid =0
for i in range(len(self.fitArray)):
if self.fitArray[i]==-1:
self.invalid = self.invalid+1
elif self.fitArray[i]!=-1:
total = total + self.fitArray[i]
avg = total/(len(self.fitArray)-self.invalid)
return avg
def evaluate(self):
for ind in range(len(self.cars)):
self.cars[ind].express()
nproc = self.nr_processes
times = []
batches = len(self.cars)/nproc+1
for batch in range(batches):
if batch is not range(batches)[-1]:
indivs = [a_+(nproc*batch) for a_ in range(nproc)]
else:
indivs = [a_+(nproc*batch) for a_ in range(len(self.cars)%nproc)]
results = pprocess.Map(limit=nproc, reuse=1)
parallel_function = results.manage(pprocess.MakeReusable(self.fitfun))
[parallel_function(self.cars[ind].phenotype, (ind-nproc*batch)) for ind in indivs];
times.extend(results[0:nproc])
time.sleep(1)
for ind in range(len(self.cars)):
times[ind] = float(times[ind])
return times
return times