def plot(self):
training_nr = 1
spiketrain1 = self.ea.best_overall_individual.spiketrain
spiketrain2 = read_training_files()
title_string = "Pop size: %s. Generations: %s Fitness: %s Distance: %s \nC-rate: %s M-rate: %s Overproduction: %s R-Max: %s Dataset: %s \n%s"%(
self.ea.population_size,
self.ea.generation,
self.ea.best_overall_individual.fitness,
self.ea.best_overall_individual.distance,
self.ea.crossover_rate,
self.ea.mutation_probability,
self.ea.overproduction_factor,
self.ea.rank_max,
training_nr,
self.ea.best_overall_individual)
fig = plt.figure()
plt.title(title_string, fontsize=10)
plt.plot(xrange(0,1001), spiketrain1, xrange(0,1001), spiketrain2)
plt.ylabel("Test spiketrain vs best evolved spiketrain")
run_string = "%s-%spop-%sgen-%sfit-%sdist-%scross-%smut-%sop-%srmax-%sdata-%s.png"%("EA",
self.ea.population_size,
self.ea.generation,
self.ea.best_overall_individual.fitness,
self.ea.best_overall_individual.distance,
self.ea.crossover_rate,
self.ea.mutation_probability,
self.ea.overproduction_factor,
self.ea.rank_max,
training_nr,
self.ea.best_overall_individual)
fig.savefig("spiketrains/"+run_string+".png")
print run_string +" fit:" +str(self.ea.best_overall_individual.fitness)+" dist:"+str(self.ea.best_overall_individual.distance) + str(self.ea.best_overall_individual)
fig = plt.figure()
plt.title(title_string)
plt.subplot(211)
plt.plot(self.generation, self.max_fitness, self.generation, self.avg_fitness)
plt.ylabel("Max and average fitness")
plt.subplot(212)
plt.plot(self.generation, self.std_deviation)
plt.ylabel("Standard deviation")
fig.savefig("fitnessplots/"+run_string+".png")
#WAR done, increment fitness of winner based on points
if points_commander>points_opponent:
commander.increment_fitness(2)
if points_commander<points_opponent:
opposing_commander.increment_fitness(2)
if points_commander==points_opponent:
commander.increment_fitness(1)
opposing_commander.increment_fitness(1)
commander.reset()
opposing_commander.reset()
#3 different fitness calculations:
#Spike Time
#Spike Interval
#Waveform
trainingdata = read_training_files()
def izzy_spike_time(p, training_spikes):
power = 4
S_b = training_spikes
S_a = p.spikes
N = min(len(S_a),len(S_b))
if N == 0:
return 0.0
sigma = 0
for t_ai, t_bi in zip(S_a, S_b):
sigma += abs(t_ai - t_bi)**power
nv = sigma ** (power ** -1)
dist = nv/N
