def crossover(protein,parents):
possible_rotations = [0,90,180,270]
is_invalid = True
while is_invalid:
is_invalid = False
p1 = choice(parents)
p2 = choice(parents)
parent1 = p1[0]
parent2 = p2[0]
offspring = []
cut_point = choice(range(len(protein)))
rotation_to_apply = choice(possible_rotations)
offspring += parent1[:cut_point+1]
pivot = offspring[cut_point]
offspring_tail = remap_parent2_coords(pivot,parent2[cut_point:],rotation_to_apply)
if offspring_tail == None:
is_invalid = True
else:
for c in offspring_tail:
if c in offspring:
is_invalid = True
break
else:
offspring += offspring_tail
new_ind = (offspring,fitness_function(protein,offspring))
return new_ind
def monte_carlo_mutation(protein,ind):
possible_rotations = [90,180,270]
mutation_tries = 0 #defines the number of tries to mutate the indivuidual
conformation_s1 = ind[0]
is_invalid = True
while is_invalid and mutation_tries != MUTATION_TRIES:
#defines the am that on wich the rotation will be applied
am_pivot = choice(range(len(conformation_s1)))
rotation_to_apply = choice(possible_rotations)
#inits the conformation_s2 as an empty list
conformation_s2 = []
for i in range(len(conformation_s1)):
is_invalid = False
if i > am_pivot:
#bring to (0,0) to rotate using am_pivot as reference
am_to_rotate = (lambda p1,p2: (p1[0] - p2[0], p1[1] - p2[1])) (conformation_s1[i], conformation_s1[am_pivot])
new_pos = apply_rotation(am_to_rotate,rotation_to_apply)
#put the point back in the position it belongs
new_pos = (lambda p1,p2: (p1[0] + p2[0], p1[1] + p2[1])) (new_pos, conformation_s1[am_pivot])
if new_pos in conformation_s2:
#we have to try another rotation
mutation_tries += 1
is_invalid = True
break
else:
conformation_s2.append(new_pos)
else:
conformation_s2.append(conformation_s1[i])
if not is_invalid:
new_ind = (conformation_s2,fitness_function(protein,conformation_s2))
return check_acceptance(new_ind,ind)
else:
return ind
population.append(temprow)
##############
# MAIN LOOP: #
##############
for idx in range(GENERATIONS):
# Run each population member and score it.
scores = []
for i in range(POPULATION_SIZE):
gene = population[i]
instance = CKBotSimEngine.CKBotSim(robotfile)
set_periodic_gait_from_GA(instance, gene, instance.gain,
free_modules)
instance.run(SIMULATION_STEPS)
fitness = fitness_function(instance, traits)
scores.append(fitness)
poses = instance.pose_info
# Write all the information to text files for post_processing.
if filename != "none":
# Write gene/fitness information.
str_out = ""
for j in range(len(gene)):
str_out = str_out + str(gene[j]) + " "
f_gene.write(str_out + "\n")
f_gene.write(str(fitness) + "\n")
# Write pose information
for j in range(len(poses)):
temprow.extend([5*random.randint(0,13), random.randint(0,5), random.randint(0,7)]) population.append(temprow) ############## # MAIN LOOP: # ############## for idx in range(GENERATIONS): # Run each population member and score it. scores = [] for i in range(POPULATION_SIZE): gene = population[i] instance = CKBotSimEngine.CKBotSim(robotfile) set_periodic_gait_from_GA(instance, gene, instance.gain, free_modules) instance.run(SIMULATION_STEPS) fitness = fitness_function(instance, traits) scores.append(fitness) poses = instance.pose_info # Write all the information to text files for post_processing. if filename != "none": # Write gene/fitness information. str_out = "" for j in range(len(gene)): str_out = str_out + str(gene[j]) + " " f_gene.write(str_out+"\n") f_gene.write(str(fitness)+"\n") # Write pose information for j in range(len(poses)):
