You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 

289 lines
25 KiB

#include "storm/modelchecker/prctl/helper/SymbolicMdpPrctlHelper.h"
#include "storm/solver/SymbolicMinMaxLinearEquationSolver.h"
#include "storm/storage/dd/DdManager.h"
#include "storm/storage/dd/Add.h"
#include "storm/storage/dd/Bdd.h"
#include "storm/utility/graph.h"
#include "storm/utility/constants.h"
#include "storm/models/symbolic/StandardRewardModel.h"
#include "storm/modelchecker/results/SymbolicQualitativeCheckResult.h"
#include "storm/modelchecker/results/SymbolicQuantitativeCheckResult.h"
#include "storm/exceptions/InvalidPropertyException.h"
#include "storm/exceptions/InvalidArgumentException.h"
#include "storm/exceptions/UncheckedRequirementException.h"
namespace storm {
namespace modelchecker {
namespace helper {
enum class EquationSystemType {
UntilProbabilities,
ExpectedRewards
};
template<storm::dd::DdType DdType, typename ValueType>
storm::dd::Bdd<DdType> computeValidSchedulerHint(EquationSystemType const& type, storm::models::symbolic::NondeterministicModel<DdType, ValueType> const& model, storm::dd::Add<DdType, ValueType> const& transitionMatrix, storm::dd::Bdd<DdType> const& maybeStates, storm::dd::Bdd<DdType> const& targetStates) {
storm::dd::Bdd<DdType> result;
if (type == EquationSystemType::UntilProbabilities) {
result = storm::utility::graph::computeSchedulerProbGreater0E(model, transitionMatrix.notZero(), maybeStates, targetStates);
} else if (type == EquationSystemType::ExpectedRewards) {
result = storm::utility::graph::computeSchedulerProb1E(model, transitionMatrix.notZero(), maybeStates, targetStates, maybeStates || targetStates);
}
return result;
}
template<storm::dd::DdType DdType, typename ValueType>
std::unique_ptr<CheckResult> SymbolicMdpPrctlHelper<DdType, ValueType>::computeUntilProbabilities(OptimizationDirection dir, storm::models::symbolic::NondeterministicModel<DdType, ValueType> const& model, storm::dd::Add<DdType, ValueType> const& transitionMatrix, storm::dd::Bdd<DdType> const& phiStates, storm::dd::Bdd<DdType> const& psiStates, bool qualitative, storm::solver::SymbolicGeneralMinMaxLinearEquationSolverFactory<DdType, ValueType> const& linearEquationSolverFactory) {
// We need to identify the states which have to be taken out of the matrix, i.e. all states that have
// probability 0 and 1 of satisfying the until-formula.
std::pair<storm::dd::Bdd<DdType>, storm::dd::Bdd<DdType>> statesWithProbability01;
if (dir == OptimizationDirection::Minimize) {
statesWithProbability01 = storm::utility::graph::performProb01Min(model, phiStates, psiStates);
} else {
statesWithProbability01 = storm::utility::graph::performProb01Max(model, phiStates, psiStates);
}
storm::dd::Bdd<DdType> maybeStates = !statesWithProbability01.first && !statesWithProbability01.second && model.getReachableStates();
STORM_LOG_INFO("Preprocessing: " << statesWithProbability01.first.getNonZeroCount() << " states with probability 1, " << statesWithProbability01.second.getNonZeroCount() << " with probability 0 (" << maybeStates.getNonZeroCount() << " states remaining).");
// Check whether we need to compute exact probabilities for some states.
if (qualitative) {
// Set the values for all maybe-states to 0.5 to indicate that their probability values are neither 0 nor 1.
return std::unique_ptr<CheckResult>(new storm::modelchecker::SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), statesWithProbability01.second.template toAdd<ValueType>() + maybeStates.template toAdd<ValueType>() * model.getManager().getConstant(storm::utility::convertNumber<ValueType>(0.5))));
} else {
// If there are maybe states, we need to solve an equation system.
if (!maybeStates.isZero()) {
// Create the matrix and the vector for the equation system.
storm::dd::Add<DdType, ValueType> maybeStatesAdd = maybeStates.template toAdd<ValueType>();
// Start by cutting away all rows that do not belong to maybe states. Note that this leaves columns targeting
// non-maybe states in the matrix.
storm::dd::Add<DdType, ValueType> submatrix = transitionMatrix * maybeStatesAdd;
// Then compute the vector that contains the one-step probabilities to a state with probability 1 for all
// maybe states.
storm::dd::Add<DdType, ValueType> prob1StatesAsColumn = statesWithProbability01.second.template toAdd<ValueType>();
prob1StatesAsColumn = prob1StatesAsColumn.swapVariables(model.getRowColumnMetaVariablePairs());
storm::dd::Add<DdType, ValueType> subvector = submatrix * prob1StatesAsColumn;
subvector = subvector.sumAbstract(model.getColumnVariables());
// Finally cut away all columns targeting non-maybe states.
submatrix *= maybeStatesAdd.swapVariables(model.getRowColumnMetaVariablePairs());
// Now solve the resulting equation system.
std::unique_ptr<storm::solver::SymbolicMinMaxLinearEquationSolver<DdType, ValueType>> solver = linearEquationSolverFactory.create(submatrix, maybeStates, model.getIllegalMask() && maybeStates, model.getRowVariables(), model.getColumnVariables(), model.getNondeterminismVariables(), model.getRowColumnMetaVariablePairs());
// If we minimize, we know that the solution is unique
if (dir == storm::solver::OptimizationDirection::Minimize) {
solver->setHasUniqueSolution();
}
// Check requirements of solver.
storm::solver::MinMaxLinearEquationSolverRequirements requirements = solver->getRequirements(dir);
boost::optional<storm::dd::Bdd<DdType>> initialScheduler;
if (!requirements.empty()) {
if (requirements.requires(storm::solver::MinMaxLinearEquationSolverRequirements::Element::ValidInitialScheduler)) {
STORM_LOG_DEBUG("Computing valid scheduler, because the solver requires it.");
initialScheduler = computeValidSchedulerHint(EquationSystemType::UntilProbabilities, model, transitionMatrix, maybeStates, statesWithProbability01.second);
requirements.clearValidInitialScheduler();
}
requirements.clearBounds();
STORM_LOG_THROW(requirements.empty(), storm::exceptions::UncheckedRequirementException, "Could not establish requirements of solver.");
}
if (initialScheduler) {
solver->setInitialScheduler(initialScheduler.get());
}
solver->setBounds(storm::utility::zero<ValueType>(), storm::utility::one<ValueType>());
solver->setRequirementsChecked();
storm::dd::Add<DdType, ValueType> result = solver->solveEquations(dir, model.getManager().template getAddZero<ValueType>(), subvector);
return std::unique_ptr<CheckResult>(new storm::modelchecker::SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), statesWithProbability01.second.template toAdd<ValueType>() + result));
} else {
return std::unique_ptr<CheckResult>(new storm::modelchecker::SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), statesWithProbability01.second.template toAdd<ValueType>()));
}
}
}
template<storm::dd::DdType DdType, typename ValueType>
std::unique_ptr<CheckResult> SymbolicMdpPrctlHelper<DdType, ValueType>::computeGloballyProbabilities(OptimizationDirection dir, storm::models::symbolic::NondeterministicModel<DdType, ValueType> const& model, storm::dd::Add<DdType, ValueType> const& transitionMatrix, storm::dd::Bdd<DdType> const& psiStates, bool qualitative, storm::solver::SymbolicGeneralMinMaxLinearEquationSolverFactory<DdType, ValueType> const& linearEquationSolverFactory) {
std::unique_ptr<CheckResult> result = computeUntilProbabilities(dir == OptimizationDirection::Minimize ? OptimizationDirection::Maximize : OptimizationDirection::Minimize, model, transitionMatrix, model.getReachableStates(), !psiStates && model.getReachableStates(), qualitative, linearEquationSolverFactory);
result->asQuantitativeCheckResult<ValueType>().oneMinus();
return result;
}
template<storm::dd::DdType DdType, typename ValueType>
std::unique_ptr<CheckResult> SymbolicMdpPrctlHelper<DdType, ValueType>::computeNextProbabilities(OptimizationDirection dir, storm::models::symbolic::NondeterministicModel<DdType, ValueType> const& model, storm::dd::Add<DdType, ValueType> const& transitionMatrix, storm::dd::Bdd<DdType> const& nextStates) {
storm::dd::Add<DdType, ValueType> result = (transitionMatrix * nextStates.swapVariables(model.getRowColumnMetaVariablePairs()).template toAdd<ValueType>()).sumAbstract(model.getColumnVariables());
if (dir == OptimizationDirection::Minimize) {
result = (result + model.getIllegalMask().template toAdd<ValueType>()).minAbstract(model.getNondeterminismVariables());
} else {
result = result.maxAbstract(model.getNondeterminismVariables());
}
return std::unique_ptr<CheckResult>(new SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), result));
}
template<storm::dd::DdType DdType, typename ValueType>
std::unique_ptr<CheckResult> SymbolicMdpPrctlHelper<DdType, ValueType>::computeBoundedUntilProbabilities(OptimizationDirection dir, storm::models::symbolic::NondeterministicModel<DdType, ValueType> const& model, storm::dd::Add<DdType, ValueType> const& transitionMatrix, storm::dd::Bdd<DdType> const& phiStates, storm::dd::Bdd<DdType> const& psiStates, uint_fast64_t stepBound, storm::solver::SymbolicGeneralMinMaxLinearEquationSolverFactory<DdType, ValueType> const& linearEquationSolverFactory) {
// We need to identify the states which have to be taken out of the matrix, i.e. all states that have
// probability 0 or 1 of satisfying the until-formula.
storm::dd::Bdd<DdType> statesWithProbabilityGreater0;
if (dir == OptimizationDirection::Minimize) {
statesWithProbabilityGreater0 = storm::utility::graph::performProbGreater0A(model, transitionMatrix.notZero(), phiStates, psiStates);
} else {
statesWithProbabilityGreater0 = storm::utility::graph::performProbGreater0E(model, transitionMatrix.notZero(), phiStates, psiStates);
}
storm::dd::Bdd<DdType> maybeStates = statesWithProbabilityGreater0 && !psiStates && model.getReachableStates();
// If there are maybe states, we need to perform matrix-vector multiplications.
if (!maybeStates.isZero()) {
// Create the matrix and the vector for the equation system.
storm::dd::Add<DdType, ValueType> maybeStatesAdd = maybeStates.template toAdd<ValueType>();
// Start by cutting away all rows that do not belong to maybe states. Note that this leaves columns targeting
// non-maybe states in the matrix.
storm::dd::Add<DdType, ValueType> submatrix = transitionMatrix * maybeStatesAdd;
// Then compute the vector that contains the one-step probabilities to a state with probability 1 for all
// maybe states.
storm::dd::Add<DdType, ValueType> prob1StatesAsColumn = psiStates.template toAdd<ValueType>().swapVariables(model.getRowColumnMetaVariablePairs());
storm::dd::Add<DdType, ValueType> subvector = (submatrix * prob1StatesAsColumn).sumAbstract(model.getColumnVariables());
// Finally cut away all columns targeting non-maybe states.
submatrix *= maybeStatesAdd.swapVariables(model.getRowColumnMetaVariablePairs());
std::unique_ptr<storm::solver::SymbolicMinMaxLinearEquationSolver<DdType, ValueType>> solver = linearEquationSolverFactory.create(submatrix, maybeStates, model.getIllegalMask() && maybeStates, model.getRowVariables(), model.getColumnVariables(), model.getNondeterminismVariables(), model.getRowColumnMetaVariablePairs());
storm::dd::Add<DdType, ValueType> result = solver->multiply(dir, model.getManager().template getAddZero<ValueType>(), &subvector, stepBound);
return std::unique_ptr<CheckResult>(new storm::modelchecker::SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), psiStates.template toAdd<ValueType>() + result));
} else {
return std::unique_ptr<CheckResult>(new storm::modelchecker::SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), psiStates.template toAdd<ValueType>()));
}
}
template<storm::dd::DdType DdType, typename ValueType>
std::unique_ptr<CheckResult> SymbolicMdpPrctlHelper<DdType, ValueType>::computeInstantaneousRewards(OptimizationDirection dir, storm::models::symbolic::NondeterministicModel<DdType, ValueType> const& model, storm::dd::Add<DdType, ValueType> const& transitionMatrix, RewardModelType const& rewardModel, uint_fast64_t stepBound, storm::solver::SymbolicGeneralMinMaxLinearEquationSolverFactory<DdType, ValueType> const& linearEquationSolverFactory) {
// Only compute the result if the model has at least one reward this->getModel().
STORM_LOG_THROW(rewardModel.hasStateRewards(), storm::exceptions::InvalidPropertyException, "Missing reward model for formula. Skipping formula.");
// Perform the matrix-vector multiplication.
std::unique_ptr<storm::solver::SymbolicMinMaxLinearEquationSolver<DdType, ValueType>> solver = linearEquationSolverFactory.create(transitionMatrix, model.getReachableStates(), model.getIllegalMask(), model.getRowVariables(), model.getColumnVariables(), model.getNondeterminismVariables(), model.getRowColumnMetaVariablePairs());
storm::dd::Add<DdType, ValueType> result = solver->multiply(dir, rewardModel.getStateRewardVector(), nullptr, stepBound);
return std::unique_ptr<CheckResult>(new SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), result));
}
template<storm::dd::DdType DdType, typename ValueType>
std::unique_ptr<CheckResult> SymbolicMdpPrctlHelper<DdType, ValueType>::computeCumulativeRewards(OptimizationDirection dir, storm::models::symbolic::NondeterministicModel<DdType, ValueType> const& model, storm::dd::Add<DdType, ValueType> const& transitionMatrix, RewardModelType const& rewardModel, uint_fast64_t stepBound, storm::solver::SymbolicGeneralMinMaxLinearEquationSolverFactory<DdType, ValueType> const& linearEquationSolverFactory) {
// Only compute the result if the model has at least one reward this->getModel().
STORM_LOG_THROW(!rewardModel.empty(), storm::exceptions::InvalidPropertyException, "Missing reward model for formula. Skipping formula.");
// Compute the reward vector to add in each step based on the available reward models.
storm::dd::Add<DdType, ValueType> totalRewardVector = rewardModel.getTotalRewardVector(transitionMatrix, model.getColumnVariables());
// Perform the matrix-vector multiplication.
std::unique_ptr<storm::solver::SymbolicMinMaxLinearEquationSolver<DdType, ValueType>> solver = linearEquationSolverFactory.create(model.getTransitionMatrix(), model.getReachableStates(), model.getIllegalMask(), model.getRowVariables(), model.getColumnVariables(), model.getNondeterminismVariables(), model.getRowColumnMetaVariablePairs());
storm::dd::Add<DdType, ValueType> result = solver->multiply(dir, model.getManager().template getAddZero<ValueType>(), &totalRewardVector, stepBound);
return std::unique_ptr<CheckResult>(new SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), result));
}
template<storm::dd::DdType DdType, typename ValueType>
std::unique_ptr<CheckResult> SymbolicMdpPrctlHelper<DdType, ValueType>::computeReachabilityRewards(OptimizationDirection dir, storm::models::symbolic::NondeterministicModel<DdType, ValueType> const& model, storm::dd::Add<DdType, ValueType> const& transitionMatrix, RewardModelType const& rewardModel, storm::dd::Bdd<DdType> const& targetStates, bool qualitative, storm::solver::SymbolicGeneralMinMaxLinearEquationSolverFactory<DdType, ValueType> const& linearEquationSolverFactory) {
// Only compute the result if there is at least one reward model.
STORM_LOG_THROW(!rewardModel.empty(), storm::exceptions::InvalidPropertyException, "Missing reward model for formula. Skipping formula.");
// Determine which states have a reward of infinity by definition.
storm::dd::Bdd<DdType> infinityStates;
storm::dd::Bdd<DdType> transitionMatrixBdd = transitionMatrix.notZero();
if (dir == OptimizationDirection::Minimize) {
infinityStates = storm::utility::graph::performProb1E(model, transitionMatrixBdd, model.getReachableStates(), targetStates, storm::utility::graph::performProbGreater0E(model, transitionMatrixBdd, model.getReachableStates(), targetStates));
} else {
infinityStates = storm::utility::graph::performProb1A(model, transitionMatrixBdd, targetStates, storm::utility::graph::performProbGreater0A(model, transitionMatrixBdd, model.getReachableStates(), targetStates));
}
infinityStates = !infinityStates && model.getReachableStates();
storm::dd::Bdd<DdType> maybeStates = (!targetStates && !infinityStates) && model.getReachableStates();
STORM_LOG_INFO("Preprocessing: " << infinityStates.getNonZeroCount() << " states with reward infinity, " << targetStates.getNonZeroCount() << " target states (" << maybeStates.getNonZeroCount() << " states remaining).");
// Check whether we need to compute exact rewards for some states.
if (qualitative) {
// Set the values for all maybe-states to 1 to indicate that their reward values
// are neither 0 nor infinity.
return std::unique_ptr<CheckResult>(new SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), infinityStates.ite(model.getManager().getConstant(storm::utility::infinity<ValueType>()), model.getManager().template getAddZero<ValueType>()) + maybeStates.template toAdd<ValueType>() * model.getManager().getConstant(storm::utility::one<ValueType>())));
} else {
// If there are maybe states, we need to solve an equation system.
if (!maybeStates.isZero()) {
// Create the matrix and the vector for the equation system.
storm::dd::Add<DdType, ValueType> maybeStatesAdd = maybeStates.template toAdd<ValueType>();
// Start by cutting away all rows that do not belong to maybe states. Note that this leaves columns targeting
// non-maybe states in the matrix.
storm::dd::Add<DdType, ValueType> submatrix = transitionMatrix * maybeStatesAdd;
// Then compute the state reward vector to use in the computation.
storm::dd::Add<DdType, ValueType> subvector = rewardModel.getTotalRewardVector(maybeStatesAdd, submatrix, model.getColumnVariables());
// Since we are cutting away target and infinity states, we need to account for this by giving
// choices the value infinity that have some successor contained in the infinity states.
storm::dd::Bdd<DdType> choicesWithInfinitySuccessor = (maybeStates && transitionMatrixBdd && infinityStates.swapVariables(model.getRowColumnMetaVariablePairs())).existsAbstract(model.getColumnVariables());
subvector = choicesWithInfinitySuccessor.ite(model.getManager().template getInfinity<ValueType>(), subvector);
// Finally cut away all columns targeting non-maybe states.
submatrix *= maybeStatesAdd.swapVariables(model.getRowColumnMetaVariablePairs());
// Now solve the resulting equation system.
std::unique_ptr<storm::solver::SymbolicMinMaxLinearEquationSolver<DdType, ValueType>> solver = linearEquationSolverFactory.create(submatrix, maybeStates, model.getIllegalMask() && maybeStates, model.getRowVariables(), model.getColumnVariables(), model.getNondeterminismVariables(), model.getRowColumnMetaVariablePairs());
// If we maximize, we know that the solution is unique
if (dir == storm::solver::OptimizationDirection::Maximize) {
solver->setHasUniqueSolution();
}
// Check requirements of solver.
storm::solver::MinMaxLinearEquationSolverRequirements requirements = solver->getRequirements(dir);
boost::optional<storm::dd::Bdd<DdType>> initialScheduler;
if (!requirements.empty()) {
if (requirements.requires(storm::solver::MinMaxLinearEquationSolverRequirements::Element::ValidInitialScheduler)) {
STORM_LOG_DEBUG("Computing valid scheduler, because the solver requires it.");
initialScheduler = computeValidSchedulerHint(EquationSystemType::ExpectedRewards, model, transitionMatrix, maybeStates, targetStates);
requirements.clearValidInitialScheduler();
}
requirements.clearLowerBounds();
STORM_LOG_THROW(requirements.empty(), storm::exceptions::UncheckedRequirementException, "Could not establish requirements of solver.");
}
if (initialScheduler) {
solver->setInitialScheduler(initialScheduler.get());
}
solver->setLowerBound(storm::utility::zero<ValueType>());
solver->setRequirementsChecked();
storm::dd::Add<DdType, ValueType> result = solver->solveEquations(dir, model.getManager().template getAddZero<ValueType>(), subvector);
return std::unique_ptr<CheckResult>(new storm::modelchecker::SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), infinityStates.ite(model.getManager().getConstant(storm::utility::infinity<ValueType>()), result)));
} else {
return std::unique_ptr<CheckResult>(new storm::modelchecker::SymbolicQuantitativeCheckResult<DdType, ValueType>(model.getReachableStates(), infinityStates.ite(model.getManager().getConstant(storm::utility::infinity<ValueType>()), model.getManager().template getAddZero<ValueType>())));
}
}
}
template class SymbolicMdpPrctlHelper<storm::dd::DdType::CUDD, double>;
template class SymbolicMdpPrctlHelper<storm::dd::DdType::Sylvan, double>;
template class SymbolicMdpPrctlHelper<storm::dd::DdType::Sylvan, storm::RationalNumber>;
}
}
}