Browse Source

More statistics output.

tempestpy_adaptions
Tim Quatmann 4 years ago
parent
commit
5213be9b69
  1. 6
      src/storm/modelchecker/multiobjective/pcaa/StandardMaPcaaWeightVectorChecker.cpp
  2. 1
      src/storm/modelchecker/multiobjective/pcaa/StandardPcaaWeightVectorChecker.cpp
  3. 29
      src/storm/modelchecker/multiobjective/preprocessing/SparseMultiObjectivePreprocessorResult.h

6
src/storm/modelchecker/multiobjective/pcaa/StandardMaPcaaWeightVectorChecker.cpp

@ -10,6 +10,8 @@
#include "storm/utility/SignalHandler.h"
#include "storm/logic/Formulas.h"
#include "storm/solver/SolverSelectionOptions.h"
#include "storm/settings/SettingsManager.h"
#include "storm/settings/modules/CoreSettings.h"
#include "storm/environment/solver/SolverEnvironment.h"
#include "storm/environment/solver/MinMaxSolverEnvironment.h"
@ -63,6 +65,10 @@ namespace storm {
STORM_LOG_WARN_COND(this->objectives[objIndex].originalFormula->isProbabilityOperatorFormula() && this->objectives[objIndex].originalFormula->asProbabilityOperatorFormula().getSubformula().isBoundedUntilFormula(), "Objective " << this->objectives[objIndex].originalFormula << " was simplified to a cumulative reward formula. Correctness of the algorithm is unknown for this type of property.");
}
}
// Print some statistics (if requested)
if (storm::settings::getModule<storm::settings::modules::CoreSettings>().isShowStatisticsSet()) {
STORM_PRINT_AND_LOG("Final preprocessed model has " << markovianStates.getNumberOfSetBits() << " Markovian states." << std::endl);
}
}
template <class SparseMdpModelType>

1
src/storm/modelchecker/multiobjective/pcaa/StandardPcaaWeightVectorChecker.cpp

@ -119,6 +119,7 @@ namespace storm {
}
STORM_PRINT_AND_LOG(numLraMecStates << " states lie on such an end component." << std::endl);
}
STORM_PRINT_AND_LOG(std::endl);
}
}

29
src/storm/modelchecker/multiobjective/preprocessing/SparseMultiObjectivePreprocessorResult.h

@ -38,23 +38,32 @@ namespace storm {
// Intentionally left empty
}
bool containsOnlyTotalRewardFormulas() const {
uint64_t getNumberOfTotalRewardFormulas() const {
uint64_t count = 0;
for (auto const& obj : objectives) {
if (!obj.formula->isRewardOperatorFormula() || !obj.formula->getSubformula().isTotalRewardFormula()) {
return false;
if (obj.formula->isRewardOperatorFormula() && obj.formula->getSubformula().isTotalRewardFormula()) {
++count;
}
}
return true;
return count;
}
bool containsLongRunAverageRewardFormulas() const {
bool containsOnlyTotalRewardFormulas() const {
return getNumberOfTotalRewardFormulas() == objectives.size();
}
uint64_t getNumberOfLongRunAverageRewardFormulas() const {
uint64_t count = 0;
for (auto const& obj : objectives) {
if (obj.formula->isRewardOperatorFormula() && obj.formula->getSubformula().isLongRunAverageRewardFormula()) {
return true;
++count;
}
}
return false;
return count;
}
bool containsLongRunAverageRewardFormulas() const {
return getNumberOfLongRunAverageRewardFormulas() > 0;
}
bool containsOnlyTrivialObjectives() const {
@ -88,12 +97,14 @@ namespace storm {
out << "--------------------------------------------------------------" << std::endl;
out << "\t" << originalFormula << std::endl;
out << std::endl;
out << "Objectives:" << std::endl;
out << "The query considers " << objectives.size() << " objectives:" << std::endl;
out << "--------------------------------------------------------------" << std::endl;
for (auto const& obj : objectives) {
obj.printToStream(out);
out << std::endl;
}
out << "Number of Long-Run-Average Reward Objectives (after preprocessing): " << getNumberOfLongRunAverageRewardFormulas() << "." << std::endl;
out << "Number of Total Reward Objectives (after preprocessing): " << getNumberOfTotalRewardFormulas() << "." << std::endl;
out << "--------------------------------------------------------------" << std::endl;
out << std::endl;
out << "Original Model Information:" << std::endl;

Loading…
Cancel
Save