diff --git a/.gitignore b/.gitignore
index 7aa5fc0a2f338b13b8bd7f9765e9a051d98a96a0..1d2761d31d0db25e70455691eb1a8e97e306c075 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,3 +1,4 @@
+BayesValidRox/.spyderproject/
tests/BeamTest/Polynomials/
tests/BeamTest/surrogateBeam/*.zip
tests/BraninTest/Polynomials/*.npy
@@ -56,6 +57,7 @@ tests/MECBM/SurrogateModel/SeqDesign_25_100_DPP/
tests/MECBM/SurrogateModel/SeqDesign_25_150_DKL/
tests/MECBM/SurrogateModel/__pycache__/read_MECBM.cpython-37.pyc
tests/MECBM/model_exe_orig.sh
+tests/Fetzer/Outputs_*
BayesValidRox/BayesInference/__pycache__/Discrepancy.cpython-36.pyc
BayesValidRox/surrogate_models/__pycache__/Exploration.cpython-36.pyc
BayesValidRox/tests/AnalyticalFunction/AnalyticFunc_Results.pkl
diff --git a/BayesValidRox/PostProcessing/PostProcessing.py b/BayesValidRox/PostProcessing/PostProcessing.py
index ff1c83604bba1e847b62138eff8787d1815fdb6a..af4a7443bf8291039d3f4756f86441a850bec60f 100644
--- a/BayesValidRox/PostProcessing/PostProcessing.py
+++ b/BayesValidRox/PostProcessing/PostProcessing.py
@@ -590,6 +590,7 @@ class PostProcessing:
"""
PCEModel = self.PCEModel
+ totalNSamples = PCEModel.ExpDesign.X.shape[0]
plotList = ['Modified LOO error', 'Validation error', 'KLD', 'BME', 'RMSEMean', 'RMSEStd']
seqList = [PCEModel.SeqModifiedLOO,PCEModel.seqValidError, PCEModel.SeqKLD, PCEModel.SeqBME,
@@ -647,7 +648,7 @@ class PostProcessing:
patch.set(facecolor=fill_color[idx])
- step = 5
+ step = PCEModel.ExpDesign.NrofNewSample if PCEModel.ExpDesign.NrofNewSample!=1 else 1
edge_color = ['red', 'blue', 'green']
fill_color = ['tan', 'cyan', 'lightgreen']
fig, ax = plt.subplots()
@@ -672,7 +673,7 @@ class PostProcessing:
plt.axhline(y=0.0, xmin=0, xmax=1, c='green', ls='--', lw=2)
# Plot each UtilFuncs
- labels = np.arange(NrofInitSamples, allErrors.shape[1]+NrofInitSamples, step)
+ labels = np.arange(NrofInitSamples, totalNSamples, step)
draw_plot(allErrors[:,::step], labels, edge_color, fill_color, idx)
@@ -708,8 +709,8 @@ class PostProcessing:
for idx, name in enumerate(nameUtil):
SeqValues = SeqDict[name]
- x_idx = np.arange(NrofInitSamples, SeqValues.shape[0]+NrofInitSamples)
-
+ step = PCEModel.ExpDesign.NrofNewSample if PCEModel.ExpDesign.NrofNewSample!=1 else 1
+ x_idx = np.arange(NrofInitSamples, totalNSamples+1, step)
if plot == 'KLD' or plot == 'BME':
# BME convergence if refBME is provided
if refBME_KLD is not None:
@@ -736,7 +737,7 @@ class PostProcessing:
plt.semilogy(x_idx, SeqValues.mean(axis=1), marker=markers[idx],
ls='--', lw=2, color=colors[idx], label=name.split("_rep",1)[0])
- plt.xticks(x_idx[::5])
+ plt.xticks(x_idx)
plt.xlabel('Number of runs')
plt.ylabel(plotLabel)
plt.legend()
diff --git a/BayesValidRox/surrogate_models/ExpDesigns.py b/BayesValidRox/surrogate_models/ExpDesigns.py
index 72e17c1bfdb36a76dea56da633796ed2d2aa2bb8..5f8904cd1d130c7d69f72d77b108aa2f13f8e6a2 100644
--- a/BayesValidRox/surrogate_models/ExpDesigns.py
+++ b/BayesValidRox/surrogate_models/ExpDesigns.py
@@ -29,6 +29,7 @@ class ExpDesigns:
self.NCandidate = 1
self.PostSnapshot = False
self.stepSnapshot = 2
+ self.UtilityFunction = 'None'
self.nReprications = 1
self.parNames = []
self.MAP = []
diff --git a/BayesValidRox/surrogate_models/surrogate_models.py b/BayesValidRox/surrogate_models/surrogate_models.py
index 3d5ff6c4abf8a6b9f24ea614b0bd5d140a6c7919..c0cad88de871f3e89255f2dd62ac58bc44232928 100644
--- a/BayesValidRox/surrogate_models/surrogate_models.py
+++ b/BayesValidRox/surrogate_models/surrogate_models.py
@@ -995,19 +995,21 @@ class aPCE:
temp = totalScore.copy()
temp[np.isnan(totalScore)] = -np.inf
sorted_idxtotalScore = np.argsort(temp)[::-1]
- bestIdx=sorted_idxtotalScore[:NrofNewSample][0]
+ bestIdx = sorted_idxtotalScore[:NrofNewSample]
# select the requested number of samples
if ExploreMethod == 'Voronoi':
- X_can = exploration.closestPoints[bestIdx]
-
- #plotter(self.ExpDesign.X, X_can, ExploreMethod, scoreExploration=None)
-
- # Calculate the maxmin score for the region of interest
- newSamples, maxminScore = exploration.getMCSamples(allCandidates=X_can)
-
- # select the requested number of samples
- Xnew = np.array([newSamples[np.argmax(maxminScore)]])
+ Xnew = np.zeros((NrofNewSample,ndim))
+ for i, idx in enumerate(bestIdx):
+ X_can = exploration.closestPoints[idx]
+
+ #plotter(self.ExpDesign.X, X_can, ExploreMethod, scoreExploration=None)
+
+ # Calculate the maxmin score for the region of interest
+ newSamples, maxminScore = exploration.getMCSamples(allCandidates=X_can)
+
+ # select the requested number of samples
+ Xnew[i] = newSamples[np.argmax(maxminScore)]
else:
Xnew = allCandidates[sorted_idxtotalScore[:NrofNewSample]]
@@ -1071,19 +1073,23 @@ class aPCE:
#print("ExploitScore:\n", ExploitScore)
- # Find the index for max in ExploreScore
- max_idx = np.where(ExploitScore == np.nanmax(ExploitScore))[0][0]
-
- # select the samples in the Vornoi cell with the highest ExploreScore
- selectedCandidates = exploration.closestPoints[max_idx]
- #print("selectedCandidates:\n", selectedCandidates)
-
- # Calculate the maxmin score for the region of interest
- newSamples, maxminScore = exploration.getMCSamples(allCandidates=selectedCandidates)
-
- # select the requested number of samples
- Xnew = np.array([newSamples[np.argmax(maxminScore)]])
+ # find an optimal point subset to add to the initial design
+ # by maximization of the utility score and taking care of NaN values
+ temp = ExploitScore.copy()
+ bestIdx = np.argsort(temp)[::-1][:NrofNewSample]
+ #bestIdx = sorted_idxtotalScore[:NrofNewSample]
+ Xnew = np.zeros((NrofNewSample,ndim))
+ for i, idx in enumerate(bestIdx):
+ X_can = exploration.closestPoints[idx]
+
+ #plotter(self.ExpDesign.X, X_can, ExploreMethod, scoreExploration=None)
+
+ # Calculate the maxmin score for the region of interest
+ newSamples, maxminScore = exploration.getMCSamples(allCandidates=X_can)
+
+ # select the requested number of samples
+ Xnew[i] = newSamples[np.argmax(maxminScore)]
elif ExploitMethod == 'alphabetic':
# ------- EXPLOITATION: ALPHABETIC -------
@@ -1778,7 +1784,7 @@ class aPCE:
for OutName in OutputName:
Scores_all.append(list(validError[OutName].values()))
ValidError = [item for sublist in Scores_all for item in sublist]
- print("\nValidError:", ValidError)
+ print("\nUpdated ValidError:", ValidError)
print("\n")
# Extract Modified LOO from Output
@@ -1798,7 +1804,7 @@ class aPCE:
# if it increases consistently stop the iterations
n_checks = 3
itrNr = Xfull.shape[0] - Xinit.shape[0]
- if itrNr > n_checks:
+ if itrNr > n_checks * self.ExpDesign.NrofNewSample:
ss = np.sign(SeqModifiedLOO - ModifiedLOO) #ss<0 error increasing
errorIncreases = np.sum(np.mean(ss[-2:], axis=1)) <= -1*n_checks
@@ -1812,7 +1818,7 @@ class aPCE:
break
else:
- prevPCEModel = PCEModel #PCEModel
+ prevPCEModel = PCEModel
# Store updated ModifiedLOO
diff --git a/BayesValidRox/tests/AnalyticalFunction/AnalyticalFunction_Test.py b/BayesValidRox/tests/AnalyticalFunction/AnalyticalFunction_Test.py
index 30f156b57d355f521fdc9c865046690b195cd024..b288801ec56976e83abead7731754500fd4531e3 100644
--- a/BayesValidRox/tests/AnalyticalFunction/AnalyticalFunction_Test.py
+++ b/BayesValidRox/tests/AnalyticalFunction/AnalyticalFunction_Test.py
@@ -107,12 +107,13 @@ if __name__ == "__main__":
# hypercube sampling of the input model or user-defined values of X and/or Y:
MetaModelOpts.addExpDesign()
- NrofInitSamples = 50
+ NrofInitSamples = 10
MetaModelOpts.ExpDesign.NrSamples = NrofInitSamples
MetaModelOpts.ExpDesign.SamplingMethod = 'MC' # 1)MC 2)LHS 3)PCM 4)LSCM 5)user
MetaModelOpts.ExpDesign.Method = 'sequential' # 1) normal 2) sequential
# Sequential experimental design (needed only for sequential ExpDesign)
+ MetaModelOpts.ExpDesign.NrofNewSample = 5
MetaModelOpts.ExpDesign.MaxNSamples = 75
MetaModelOpts.ExpDesign.ModifiedLOOThreshold = 1e-6
diff --git a/BayesValidRox/tests/Co2BenchmarkTest/Test_CO2Benchmark.py b/BayesValidRox/tests/Co2BenchmarkTest/Test_CO2Benchmark.py
index f0b1a250f35bd1515e196242343fe8dfcc7a240c..e30bf4ec24703d68ece3f5f21696bafb1b62f556 100644
--- a/BayesValidRox/tests/Co2BenchmarkTest/Test_CO2Benchmark.py
+++ b/BayesValidRox/tests/Co2BenchmarkTest/Test_CO2Benchmark.py
@@ -115,6 +115,7 @@ if __name__ == "__main__":
#MetaModelOpts.ExpDesign.Y = ExpDesign_Y
# Sequential experimental design (needed only for sequential ExpDesign)
+ MetaModelOpts.ExpDesign.NrofNewSample = 10
MetaModelOpts.ExpDesign.MaxNSamples = 500
MetaModelOpts.ExpDesign.ModifiedLOOThreshold = 1e-4