from functions import genesis
from functions import ras_calc
from functions import fitness
from functions import mate
from agent import agent
#Global Variables
g = 0 #instantaenous generation count
target_ras = 0 #global minima for convergence criteria
max_gen = 50 #generation for exit criteria
size = 100 #population size
mut_prob = 0.1 #overall mutation probability
fittest = []
#generation 0
population = genesis(size)
population = ras_calc(population)
while g <= max_gen: #exit criteria
#evaluation
parents = fitness(population)
fittest.append(parents[0])
ras_fittest = parents[0].ras
# convergence criteria
if ras_fittest <= target_ras:
break
#generation update (crossover,mutation,etc)
g += 1
from functions import genesis
from functions import write_block, change_block
import os
from_who = 'ivan'
to_whom = 'petr'
amount = 1002
def create_dir(num):
if not os.path.exists('{}'.format(num)):
os.makedirs('{}'.format(num))
return ('Done!')
#write_block(from_who,to_whom,amount)
genesis()
#create_dir(2)
def GA_FineMut(ind_op, gen_op, mut_prob_op, n_runs, n_genes_op):
config_perf_FineMut = {
'Config #': [],
'Mut_Prob': [],
'Fit_Ave_Fit': [],
'Ave_Ave_Fit': [],
'Weak_Ave_Fit': [],
}
raw_data_FineMut = {}
#Test config modelling
mut_configs = np.linspace(mut_prob_op - 0.05, mut_prob_op, 11)
counter = 0
for m_config in mut_configs:
counter += 1
raw_data_FineMut[f"Config # {counter}"] = []
#config status
print(f"Config # {counter}")
config_perf_FineMut['Config #'].append(counter)
config_perf_FineMut['Mut_Prob'].append(m_config)
#intermediary Data Storage Lists
percent_improvement = []
ave_cost_fittest_g = []
ave_cost_fittest_run = []
stdv_g1 = []
stdv_gEnd = []
ch_stdv = []
run = 0
while run < n_runs:
#iterator
run += 1
#GA Parameters
gen = 1
target_cost = 0
max_gen = gen_op
size = ind_op
mut_prob = m_config
mut_type = 'n_gene' #mutation type - all,gene,motor
n_genes = n_genes_op
fittest = []
search_limit = 5000
#Simulation Parameters
dt = 5e-4
tsteps = 100
#Generation 0
population = genesis(size)
#population = response(population,dt,tsteps)
#population = cost_calc(gen,population)
population = rastrigin(population)
cost_g1_s = []
for agent in population:
cost_g1_s.append(agent.ras)
#storage for plotting
cost_fittest_g = []
gen_s = []
#GA loop
while gen <= max_gen:
gen_s.append(gen)
#parents = fitness(population)
parents = fitness_ras(population)
fittest.append(parents[0])
for item in range(3):
random.shuffle(parents)
#cost_fittest = parents[0].j_fluc
cost_fittest = parents[0].ras
#cost_fittest_g.append(parents[0].j_fluc)
cost_fittest_g.append(cost_fittest)
if cost_fittest <= target_cost:
break
gen += 1
#children = mate(parents, mut_prob, mut_type, gen, max_gen, search_limit,n_genes)
children = mate_ras(parents, mut_prob, mut_type, gen, max_gen,
search_limit, n_genes)
if children == None:
break
#children = response(children,dt,tsteps)
#children = cost_calc(gen,children)
children = rastrigin(children)
parents.extend(children)
population = parents
# final stats:
'''print(f'Defined Parameters:')
print(f' Target Objective Value: {target_cost}')
print(f' Maximum Generation Count: {max_gen}')
print(f' Population Size: {size}')
print(f' Mutation Probability: {mut_prob*100} %')
print('')
if gen < max_gen:
print(f'Premature search stop at generation {gen}')
else:
print(f'Search completed for defined generations')
#minimization = fittest[0].j_fluc - fittest[-1].j_fluc
minimization = fittest[0].ras - fittest[-1].ras
#percent_improvement = (minimization/fittest[0].j_fluc)*100
percent_improvement = abs((minimization/fittest[0].ras)*100)
print(f'GA minimized: {minimization}')
print(f'Percent Improvement: {percent_improvement} %')
#print(f'Best objective value: {fittest[-1].j_fluc}')
print(f'Best objective value: {fittest[-1].ras}')
print(f'With genes: {fittest[-1].genes}')'''
cost_fittest_run = []
raw_data_FineMut[f"Config # {counter}"].append(fittest[-1].ras)
cost_fittest_run.append(fittest[-1].ras)
cost_gEnd_s = []
for agent in population:
cost_gEnd_s.append(agent.ras)
#Intermediary Performance Data
percent_improvement.append(
abs(((fittest[0].ras - fittest[-1].ras) / fittest[0].ras) *
100))
ave_cost_fittest_g.append(stat.mean(cost_fittest_g))
ave_cost_fittest_run.append(stat.mean(cost_fittest_run))
stdv_g1.append(stat.pstdev(cost_g1_s))
stdv_gEnd.append(stat.pstdev(cost_gEnd_s))
ch_stdv.append(
abs((stat.pstdev(cost_gEnd_s) - stat.pstdev(cost_g1_s)) /
stat.pstdev(cost_g1_s)))
#Final Performance Data
#config_perf['Min_%Imp'].append(min(percent_improvement))
#config_perf['Ave_%Imp'].append(stat.mean(percent_improvement))
#config_perf['Max_%Imp'].append(max(percent_improvement))
config_perf_FineMut['Fit_Ave_Fit'].append(min(ave_cost_fittest_run))
config_perf_FineMut['Ave_Ave_Fit'].append(
stat.mean(ave_cost_fittest_run))
config_perf_FineMut['Weak_Ave_Fit'].append(max(ave_cost_fittest_run))
#config_perf['Min_StDv_g1'].append(min(stdv_g1))
#config_perf['Ave_StDv_g1'].append(stat.mean(stdv_g1))
#config_perf['Max_StDv_g1'].append(max(stdv_g1))
#config_perf['Min_StDv_gEnd'].append(min(stdv_gEnd))
#config_perf['Ave_StDv_gEnd'].append(stat.mean(stdv_gEnd))
#config_perf['Max_StDv_gEnd'].append(max(stdv_gEnd))
#config_perf['Min_%Ch_StDv'].append(min(ch_stdv))
#config_perf['Ave_%Ch_StDv'].append(stat.mean(ch_stdv))
#config_perf['Max_%Ch_StDv'].append(max(ch_stdv))
return config_perf_FineMut, raw_data_FineMut
