def pareto_surface_from_kde_sampling():
from scipy.stats import uniform, gaussian_kde
n_design_variables = 1
n_obj_functions = 2
N_iterations = 10
N_samples_per_iteration = 10000
column_names = ['id'] \
+ ['x{}'.format(i+1) for i in range(n_design_variables)] \
+ ['f{}'.format(i+1) for i in range(n_obj_functions)]
data = []
df = None
for i_iteration in range(N_iterations):
if i_iteration == 0:
x_sampler = [uniform(-5,10) for i in range(n_design_variables)]
for i in range(N_samples_per_iteration):
x = [u.rvs() for u in x_sampler]
f1,f2 = schaffer(x)
row = [i] + x + [f1,f2]
data.append([i] + x + [f1,f2])
df = pd.DataFrame(data, columns=column_names)
else:
pareto_rows = pareto.pareto(df[['f1','f2']].values.tolist())
print("i={},n_pareto={}".format(i_iteration,len(pareto_rows)))
print(df.loc[pareto_rows,['x1','x2','x3']].T)
kde = gaussian_kde(df.loc[pareto_rows,['x1','x2','x3']].T)
data = []
for i in range(N_samples_per_iteration):
x = kde.resample(1)
f1,f2 = schaffer([x[i] for i in range(3)])
row = [i] + x + [f1,f2]
data.append([i] + [x[i][0] for i in range(3)] + [f1,f2])
df = pd.DataFrame(data, columns=column_names)
return df
def filter_by_pareto_membership(self,df=None,v=None):
""" filter by pareto membership
Calculation of pareto membership by removal of dominated points
Args:
df(pandas.DataFrame): the dataframe of results. By default this is set to None,
which uses the results_df property from a PyposmatDataFile which was
read into this instance.
Returns:
(tuple): the results of the pareto membership consisting of a :obj:`'list` of integers
which hare the indices of the of the surviving candidate parameterizations, and an
:obj:`OrderedDict` containing the information for Pareto membership
"""
assert df is None or isintance(df,pd.DataFrame)
if df is None:
df = self.results_df
for qn in self.qoi_names:
df['{}.abserr'.format(qn)] = df['{}.err'.format(qn)].abs()
abs_error_names = ['{}.abserr'.format(v) for v in self.qoi_names]
is_survive_idx = pareto(df[abs_error_names].values.tolist())
n_potentials_start, n_cols = df.shape
n_potentials_final = len(is_survive_idx)
pct_potentials_final = n_potentials_final/n_potentials_start
pareto_set_info = OrderedDict()
pareto_set_info['n_potentials_start'] = n_potentials_start
pareto_set_info['n_potentials_final'] = n_potentials_final
pareto_set_info['pct_potentials_final'] = pct_potentials_final
return set(is_survive_idx), pareto_set_info
datafile.read(filename=_data_fn)
from pypospack.pareto import pareto
df = copy.deepcopy(datafile.df)
nr,nc = df.shape
_nsimulations = OrderedDict()
_nsimulations['start'] = nr
abs_error_names = ["{}.abserr".format(q) for q in datafile.qoi_names]
for q in datafile.qoi_names:
qe = "{}.err".format(q)
qne = "{}.abserr".format(q)
df[qne] = df[qe].abs()
names = list(df.columns.values)
abs_err_idx = [names.index(n) for n in abs_error_names]
pareto_idx = pareto(df[abs_error_names].values.tolist())
datafile.df = df.loc[pareto_idx,datafile.names]
datafile.write("results.pareto.out")
#pareto_idx = pareto_bruteforce(df[abs_error_names].values.tolist())
#print(pareto_set)
if make_rugplots:
datafile = PyposmatDataFile()
datafile.read("results.pareto.out")
datafile.qoi_references = OrderedDict()
datafile.qoi_references['TARGET'] = copy.deepcopy(qoi_targets)
datafile.score_by_d_metric(scaling_factors='TARGET')
datafile.subselect_by_score(
score_name='d_metric',
n=_n_potentials)
data = []
for i in range(N_samples_per_iteration):
x = kde.resample(1)
f1,f2 = schaffer([x[i] for i in range(3)])
row = [i] + x + [f1,f2]
data.append([i] + [x[i][0] for i in range(3)] + [f1,f2])
df = pd.DataFrame(data, columns=column_names)
return df
if True:
plt.close("all")
fig, ax = plt.subplots(1,1)
n_simulations = 0
cg_df = pareto_surface_from_conjugate_gradient_method()
cg_pareto_rows = pareto.pareto(cg_df[['f1','f2']].values.tolist())
cg_pareto_df = cg_df.loc[cg_pareto_rows]
cg_pareto_df = cg_pareto_df.sort_values('f1')
ax.scatter(
cg_df['f1'],
cg_df['f2'],
c='black',marker=".",s=1,
)
ax.plot(
cg_pareto_df['f1'],
cg_pareto_df['f2'],
c='red',linewidth=1,marker=".",ms=5
)
#ax.set_xlim(-20,0)
#ax.set_ylim(-12.5,7.5)
ax.set_xlabel(r"$f_1(x)$")
from pypospack.pareto import pareto
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from pareto_visualization import make_pareto_plot_3d
if __name__ == "__main__":
x = np.random.normal(0,1,1000)
y = np.random.normal(0,1,1000)
z = np.random.normal(0,1,1000)
df = pd.DataFrame()
df['x'] = x
df['y'] = y
df['z'] = z
pareto_idx = pareto(df.values.tolist())
make_pareto_plot_3d(df,pareto_idx)
