Browse Source

LRA finally working for ctmcs

Former-commit-id: 699e4714a4
tempestpy_adaptions
dehnert 9 years ago
parent
commit
1e5398c8b7
  1. 194
      src/modelchecker/csl/SparseCtmcCslModelChecker.cpp
  2. 6
      src/modelchecker/csl/SparseCtmcCslModelChecker.h
  3. 271
      src/modelchecker/prctl/SparseDtmcPrctlModelChecker.cpp
  4. 2
      src/modelchecker/prctl/SparseDtmcPrctlModelChecker.h
  5. 36
      src/settings/modules/GmmxxEquationSolverSettings.cpp
  6. 34
      src/settings/modules/GmmxxEquationSolverSettings.h
  7. 19
      src/settings/modules/NativeEquationSolverSettings.cpp
  8. 18
      src/settings/modules/NativeEquationSolverSettings.h
  9. 24
      src/solver/GmmxxLinearEquationSolver.cpp
  10. 131
      src/solver/NativeLinearEquationSolver.cpp
  11. 5
      src/solver/NativeLinearEquationSolver.h
  12. 261
      src/storage/SparseMatrix.cpp
  13. 9
      src/storage/SparseMatrix.h
  14. 21
      src/utility/solver.cpp
  15. 9
      src/utility/solver.h
  16. 9
      test/functional/modelchecker/NativeDtmcPrctlModelCheckerTest.cpp

194
src/modelchecker/csl/SparseCtmcCslModelChecker.cpp

@ -12,6 +12,8 @@
#include "src/modelchecker/results/ExplicitQualitativeCheckResult.h"
#include "src/modelchecker/results/ExplicitQuantitativeCheckResult.h"
#include "src/storage/StronglyConnectedComponentDecomposition.h"
#include "src/exceptions/InvalidStateException.h"
#include "src/exceptions/InvalidPropertyException.h"
#include "src/exceptions/NotImplementedException.h"
@ -476,25 +478,193 @@ namespace storm {
std::unique_ptr<CheckResult> subResultPointer = this->check(stateFormula);
ExplicitQualitativeCheckResult const& subResult = subResultPointer->asExplicitQualitativeCheckResult();
// Since we use the uniformized matrix to compute the steady state probabilities, we need to build it first.
ValueType uniformizationRate = 0;
for (auto const& rate : this->getModel().getExitRateVector()) {
uniformizationRate = std::max(uniformizationRate, rate);
storm::storage::SparseMatrix<ValueType> probabilityMatrix = this->computeProbabilityMatrix(this->getModel().getTransitionMatrix(), this->getModel().getExitRateVector());
return std::unique_ptr<CheckResult>(new ExplicitQuantitativeCheckResult<ValueType>(computeLongRunAverageHelper(this->getModel(), probabilityMatrix, subResult.getTruthValuesVector(), &this->getModel().getExitRateVector(), qualitative, *linearEquationSolverFactory)));
}
template<typename ValueType>
std::vector<ValueType> SparseCtmcCslModelChecker<ValueType>::computeLongRunAverageHelper(storm::models::sparse::DeterministicModel<ValueType> const& model, storm::storage::SparseMatrix<ValueType> const& transitionMatrix, storm::storage::BitVector const& psiStates, std::vector<ValueType> const* exitRateVector, bool qualitative, storm::utility::solver::LinearEquationSolverFactory<ValueType> const& linearEquationSolverFactory) {
// If there are no goal states, we avoid the computation and directly return zero.
uint_fast64_t numOfStates = transitionMatrix.getRowCount();
if (psiStates.empty()) {
return std::vector<ValueType>(numOfStates, storm::utility::zero<ValueType>());
}
uniformizationRate *= 1.02;
STORM_LOG_THROW(uniformizationRate > 0, storm::exceptions::InvalidStateException, "The uniformization rate must be positive.");
// storm::storage::SparseMatrix<ValueType> uniformizedMatrix = this->computeUniformizedMatrix(this->getModel().getTransitionMatrix(), storm::storage::BitVector(this->getModel().getNumberOfStates(), true), uniformizationRate, this->getModel().getExitRateVector());
// storm::storage::SparseMatrix<ValueType> uniformizedMatrix = this->computeGeneratorMatrix(this->getModel().getTransitionMatrix(), this->getModel().getExitRateVector());
storm::storage::SparseMatrix<ValueType> uniformizedMatrix = this->computeProbabilityMatrix(this->getModel().getTransitionMatrix(), this->getModel().getExitRateVector());
// Likewise, if all bits are set, we can avoid the computation.
if (psiStates.full()) {
return std::vector<ValueType>(numOfStates, storm::utility::one<ValueType>());
}
// Start by decomposing the DTMC into its BSCCs.
storm::storage::StronglyConnectedComponentDecomposition<double> bsccDecomposition(model, false, true);
// Get some data members for convenience.
ValueType one = storm::utility::one<ValueType>();
ValueType zero = storm::utility::zero<ValueType>();
std::vector<ValueType> result = SparseDtmcPrctlModelChecker<ValueType>::computeLongRunAverageHelper(this->getModel(), uniformizedMatrix, subResult.getTruthValuesVector(), qualitative, *linearEquationSolverFactory);
// First we check which states are in BSCCs.
storm::storage::BitVector statesInBsccs(numOfStates);
storm::storage::BitVector firstStatesInBsccs(numOfStates);
std::vector<uint_fast64_t> stateToBsccIndexMap(transitionMatrix.getColumnCount());
return std::unique_ptr<CheckResult>(new ExplicitQuantitativeCheckResult<ValueType>());
for (uint_fast64_t currentBsccIndex = 0; currentBsccIndex < bsccDecomposition.size(); ++currentBsccIndex) {
storm::storage::StronglyConnectedComponent const& bscc = bsccDecomposition[currentBsccIndex];
// Gather information for later use.
bool first = true;
for (auto const& state : bscc) {
statesInBsccs.set(state);
if (first) {
firstStatesInBsccs.set(state);
}
stateToBsccIndexMap[state] = currentBsccIndex;
first = false;
}
}
firstStatesInBsccs = firstStatesInBsccs % statesInBsccs;
storm::storage::BitVector statesNotInBsccs = ~statesInBsccs;
// Then we construct an equation system that yields the steady state probabilities for all states in BSCCs.
storm::storage::SparseMatrix<ValueType> bsccEquationSystem = transitionMatrix.getSubmatrix(false, statesInBsccs, statesInBsccs, true);
// Since in the fix point equation, we need to multiply the vector from the left, we convert this to a
// multiplication from the right by transposing the system.
bsccEquationSystem = bsccEquationSystem.transpose(false, true);
// Create an auxiliary structure that makes it easy to look up the indices within the set of BSCC states.
uint_fast64_t lastIndex = 0;
uint_fast64_t currentNumberOfSetBits = 0;
std::vector<uint_fast64_t> indexInStatesInBsccs;
indexInStatesInBsccs.reserve(transitionMatrix.getRowCount());
for (auto index : statesInBsccs) {
while (lastIndex <= index) {
indexInStatesInBsccs.push_back(currentNumberOfSetBits);
++lastIndex;
}
++currentNumberOfSetBits;
}
// Now build the final equation system matrix.
storm::storage::SparseMatrixBuilder<ValueType> builder;
uint_fast64_t currentBsccIndex = 0;
for (uint_fast64_t row = 0; row < bsccEquationSystem.getRowCount(); ++row) {
// If the current row is the first one belonging to a BSCC, we substitute it by the constraint that the
// values for states of this BSCC must sum to one.
if (firstStatesInBsccs.get(row)) {
storm::storage::StronglyConnectedComponent const& bscc = bsccDecomposition[currentBsccIndex];
for (auto const& state : bscc) {
builder.addNextValue(row, indexInStatesInBsccs[state], one);
}
++currentBsccIndex;
} else {
// Otherwise, we copy the row, and subtract 1 from the diagonal.
for (auto& entry : bsccEquationSystem.getRow(row)) {
if (entry.getColumn() == row) {
builder.addNextValue(row, entry.getColumn(), entry.getValue() - one);
} else {
builder.addNextValue(row, entry.getColumn(), entry.getValue());
}
}
}
}
bsccEquationSystem = builder.build();
std::vector<ValueType> bsccEquationSystemRightSide(bsccEquationSystem.getColumnCount(), zero);
storm::utility::vector::setVectorValues(bsccEquationSystemRightSide, firstStatesInBsccs, one);
// As an initial guess, we take the uniform distribution over all states of the containing BSCC.
std::vector<ValueType> bsccEquationSystemSolution(bsccEquationSystem.getColumnCount(), zero);
for (uint_fast64_t currentBsccIndex = 0; currentBsccIndex < bsccDecomposition.size(); ++currentBsccIndex) {
storm::storage::StronglyConnectedComponent const& bscc = bsccDecomposition[currentBsccIndex];
for (auto const& state : bscc) {
bsccEquationSystemSolution[indexInStatesInBsccs[state]] = one / bscc.size();
}
}
{
std::unique_ptr<storm::solver::LinearEquationSolver<ValueType>> solver = linearEquationSolverFactory.create(bsccEquationSystem);
solver->solveEquationSystem(bsccEquationSystemSolution, bsccEquationSystemRightSide);
}
// If exit rates were given, we need to 'fix' the results to also account for the timing behaviour.
if (exitRateVector != nullptr) {
std::vector<ValueType> bsccTotalValue(bsccDecomposition.size(), zero);
std::size_t i = 0;
for (auto stateIter = statesInBsccs.begin(); stateIter != statesInBsccs.end(); ++i, ++stateIter) {
bsccTotalValue[stateToBsccIndexMap[*stateIter]] += bsccEquationSystemSolution[i] * (one / (*exitRateVector)[*stateIter]);
}
i = 0;
for (auto stateIter = statesInBsccs.begin(); stateIter != statesInBsccs.end(); ++i, ++stateIter) {
bsccEquationSystemSolution[i] = (bsccEquationSystemSolution[i] * (one / (*exitRateVector)[*stateIter])) / bsccTotalValue[stateToBsccIndexMap[*stateIter]];
}
}
// Calculate LRA Value for each BSCC from steady state distribution in BSCCs.
std::vector<ValueType> bsccLra(bsccDecomposition.size(), zero);
size_t i = 0;
for (auto stateIter = statesInBsccs.begin(); stateIter != statesInBsccs.end(); ++i, ++stateIter) {
if (psiStates.get(*stateIter)) {
bsccLra[stateToBsccIndexMap[*stateIter]] += bsccEquationSystemSolution[i];
}
}
std::vector<ValueType> rewardSolution;
if (!statesNotInBsccs.empty()) {
// Calculate LRA for states not in bsccs as expected reachability rewards.
// Target states are states in bsccs, transition reward is the lra of the bscc for each transition into a
// bscc and 0 otherwise. This corresponds to the sum of LRAs in BSCC weighted by the reachability probability
// of the BSCC.
std::vector<ValueType> rewardRightSide;
rewardRightSide.reserve(statesNotInBsccs.getNumberOfSetBits());
for (auto state : statesNotInBsccs) {
ValueType reward = zero;
for (auto entry : transitionMatrix.getRow(state)) {
if (statesInBsccs.get(entry.getColumn())) {
reward += entry.getValue() * bsccLra[stateToBsccIndexMap[entry.getColumn()]];
}
}
rewardRightSide.push_back(reward);
}
storm::storage::SparseMatrix<ValueType> rewardEquationSystemMatrix = transitionMatrix.getSubmatrix(false, statesNotInBsccs, statesNotInBsccs, true);
rewardEquationSystemMatrix.convertToEquationSystem();
rewardSolution = std::vector<ValueType>(rewardEquationSystemMatrix.getColumnCount(), one);
{
std::unique_ptr<storm::solver::LinearEquationSolver<ValueType>> solver = linearEquationSolverFactory.create(rewardEquationSystemMatrix);
solver->solveEquationSystem(rewardSolution, rewardRightSide);
}
}
// Fill the result vector.
std::vector<ValueType> result(numOfStates);
auto rewardSolutionIter = rewardSolution.begin();
for (size_t state = 0; state < numOfStates; ++state) {
if (statesInBsccs.get(state)) {
// Assign the value of the bscc the state is in.
result[state] = bsccLra[stateToBsccIndexMap[state]];
} else {
STORM_LOG_ASSERT(rewardSolutionIter != rewardSolution.end(), "Too few elements in solution.");
// Take the value from the reward computation. Since the n-th state not in any bscc is the n-th
// entry in rewardSolution we can just take the next value from the iterator.
result[state] = *rewardSolutionIter;
++rewardSolutionIter;
}
}
return result;
}
// Explicitly instantiate the model checker.
template class SparseCtmcCslModelChecker<double>;

6
src/modelchecker/csl/SparseCtmcCslModelChecker.h

@ -13,10 +13,14 @@ namespace storm {
template<storm::dd::DdType DdType, typename ValueType>
class HybridCtmcCslModelChecker;
template<typename ValueType>
class SparseDtmcPrctlModelChecker;
template<class ValueType>
class SparseCtmcCslModelChecker : public SparsePropositionalModelChecker<ValueType> {
public:
friend class HybridCtmcCslModelChecker<storm::dd::DdType::CUDD, ValueType>;
friend class SparseDtmcPrctlModelChecker<ValueType>;
explicit SparseCtmcCslModelChecker(storm::models::sparse::Ctmc<ValueType> const& model);
explicit SparseCtmcCslModelChecker(storm::models::sparse::Ctmc<ValueType> const& model, std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<ValueType>>&& linearEquationSolverFactory);
@ -40,7 +44,7 @@ namespace storm {
static std::vector<ValueType> computeUntilProbabilitiesHelper(storm::storage::SparseMatrix<ValueType> const& transitionMatrix, storm::storage::SparseMatrix<ValueType> const& backwardTransitions, storm::storage::BitVector const& phiStates, storm::storage::BitVector const& psiStates, bool qualitative, storm::utility::solver::LinearEquationSolverFactory<ValueType> const& linearEquationSolverFactory);
std::vector<ValueType> computeInstantaneousRewardsHelper(double timeBound) const;
std::vector<ValueType> computeCumulativeRewardsHelper(double timeBound) const;
static std::vector<ValueType> computeLongRunAverageHelper(storm::models::sparse::DeterministicModel<ValueType> const& model, storm::storage::SparseMatrix<ValueType> const& transitionMatrix, storm::storage::BitVector const& psiStates, std::vector<ValueType> const* exitRateVector, bool qualitative, storm::utility::solver::LinearEquationSolverFactory<ValueType> const& linearEquationSolverFactory);
/*!
* Computes the matrix representing the transitions of the uniformized CTMC.
*

271
src/modelchecker/prctl/SparseDtmcPrctlModelChecker.cpp

@ -8,11 +8,11 @@
#include "src/utility/graph.h"
#include "src/utility/solver.h"
#include "src/modelchecker/csl/SparseCtmcCslModelChecker.h"
#include "src/modelchecker/results/ExplicitQualitativeCheckResult.h"
#include "src/modelchecker/results/ExplicitQuantitativeCheckResult.h"
#include "src/exceptions/InvalidStateException.h"
#include "src/storage/StronglyConnectedComponentDecomposition.h"
#include "src/exceptions/InvalidPropertyException.h"
@ -302,280 +302,13 @@ namespace storm {
ExplicitQualitativeCheckResult const& subResult = subResultPointer->asExplicitQualitativeCheckResult();
return std::unique_ptr<CheckResult>(new ExplicitQuantitativeCheckResult<ValueType>(this->computeReachabilityRewardsHelper(this->getModel().getTransitionMatrix(), this->getModel().getOptionalStateRewardVector(), this->getModel().getOptionalTransitionRewardMatrix(), this->getModel().getBackwardTransitions(), subResult.getTruthValuesVector(), *this->linearEquationSolverFactory, qualitative)));
}
template<typename ValueType>
std::vector<ValueType> SparseDtmcPrctlModelChecker<ValueType>::computeLongRunAverageHelper(storm::models::sparse::DeterministicModel<ValueType> const& model, storm::storage::SparseMatrix<ValueType> const& transitionMatrix, storm::storage::BitVector const& psiStates, bool qualitative, storm::utility::solver::LinearEquationSolverFactory<ValueType> const& linearEquationSolverFactory) {
// If there are no goal states, we avoid the computation and directly return zero.
uint_fast64_t numOfStates = transitionMatrix.getRowCount();
if (psiStates.empty()) {
return std::vector<ValueType>(numOfStates, storm::utility::zero<ValueType>());
}
// Likewise, if all bits are set, we can avoid the computation.
if (psiStates.full()) {
return std::vector<ValueType>(numOfStates, storm::utility::one<ValueType>());
}
// Start by decomposing the DTMC into its BSCCs.
storm::storage::StronglyConnectedComponentDecomposition<double> bsccDecomposition(model, false, true);
// Get some data members for convenience.
ValueType one = storm::utility::one<ValueType>();
ValueType zero = storm::utility::zero<ValueType>();
// First we check which states are in BSCCs. We use this later to speed up reachability analysis
storm::storage::BitVector statesInBsccs(numOfStates);
storm::storage::BitVector statesInBsccsWithoutFirst(numOfStates);
std::vector<uint_fast64_t> stateToBsccIndexMap(transitionMatrix.getColumnCount());
for (uint_fast64_t currentBsccIndex = 0; currentBsccIndex < bsccDecomposition.size(); ++currentBsccIndex) {
storm::storage::StronglyConnectedComponent const& bscc = bsccDecomposition[currentBsccIndex];
// Gather information for later use.
bool first = true;
for (auto const& state : bscc) {
statesInBsccs.set(state);
if (!first) {
statesInBsccsWithoutFirst.set(state);
}
stateToBsccIndexMap[state] = currentBsccIndex;
first = false;
}
}
storm::storage::BitVector statesNotInBsccs = ~statesInBsccs;
std::cout << transitionMatrix << std::endl;
// Calculate steady state distribution for all BSCCs by calculating an eigenvector for the eigenvalue 1 of
// the transposed transition matrix for the BSCCs.
storm::storage::SparseMatrix<ValueType> bsccEquationSystem = transitionMatrix.getSubmatrix(false, statesInBsccs, statesInBsccs, true);
std::cout << bsccEquationSystem << std::endl;
bsccEquationSystem = bsccEquationSystem.transpose(false, true);
std::cout << bsccEquationSystem << std::endl;
// Subtract identity matrix.
for (uint_fast64_t row = 0; row < bsccEquationSystem.getRowCount(); ++row) {
for (auto& entry : bsccEquationSystem.getRow(row)) {
if (entry.getColumn() == row) {
entry.setValue(entry.getValue() - one);
}
}
}
std::cout << bsccEquationSystem << std::endl;
std::cout << statesInBsccsWithoutFirst << " // " << statesInBsccs << std::endl;
bsccEquationSystem = bsccEquationSystem.getSubmatrix(false, statesInBsccsWithoutFirst, statesInBsccs, false);
std::cout << bsccEquationSystem << std::endl;
// For each BSCC, add a row to the matrix that expresses that the sum over all entries in this BSCC needs to be one.
storm::storage::SparseMatrixBuilder<ValueType> builder(std::move(bsccEquationSystem));
typename storm::storage::SparseMatrixBuilder<ValueType>::index_type row = builder.getLastRow() + 1;
for (uint_fast64_t currentBsccIndex = 0; currentBsccIndex < bsccDecomposition.size(); ++currentBsccIndex) {
storm::storage::StronglyConnectedComponent const& bscc = bsccDecomposition[currentBsccIndex];
for (auto const& state : bscc) {
builder.addNextValue(row, state, one);
}
++row;
}
bsccEquationSystem = builder.build();
std::cout << bsccEquationSystem << std::endl;
std::vector<ValueType> bsccEquationSystemRightSide(bsccEquationSystem.getColumnCount(), zero);
bsccEquationSystemRightSide.back() = one;
std::vector<ValueType> bsccEquationSystemSolution(bsccEquationSystem.getColumnCount(), one);
{
std::unique_ptr<storm::solver::LinearEquationSolver<ValueType>> solver = linearEquationSolverFactory.create(bsccEquationSystem);
solver->solveEquationSystem(bsccEquationSystemSolution, bsccEquationSystemRightSide);
}
ValueType sum = zero;
for (auto const& elem : bsccEquationSystemSolution) {
std::cout << "sol " << elem << std::endl;
sum += elem;
}
std::cout << "sum: " << sum << std::endl;
std::cout << "in " << bsccDecomposition.size() << "bsccs" << std::endl;
// Calculate LRA Value for each BSCC from steady state distribution in BSCCs.
// We have to scale the results, as the probabilities for each BSCC have to sum up to one, which they don't
// necessarily do in the solution of the equation system.
std::vector<ValueType> bsccLra(bsccDecomposition.size(), zero);
std::vector<ValueType> bsccTotalValue(bsccDecomposition.size(), zero);
size_t i = 0;
for (auto stateIter = statesInBsccs.begin(); stateIter != statesInBsccs.end(); ++i, ++stateIter) {
if (psiStates.get(*stateIter)) {
bsccLra[stateToBsccIndexMap[*stateIter]] += bsccEquationSystemSolution[i];
}
bsccTotalValue[stateToBsccIndexMap[*stateIter]] += bsccEquationSystemSolution[i];
}
for (i = 0; i < bsccLra.size(); ++i) {
bsccLra[i] /= bsccTotalValue[i];
}
std::vector<ValueType> rewardSolution;
if (!statesNotInBsccs.empty()) {
// Calculate LRA for states not in bsccs as expected reachability rewards.
// Target states are states in bsccs, transition reward is the lra of the bscc for each transition into a
// bscc and 0 otherwise. This corresponds to the sum of LRAs in BSCC weighted by the reachability probability
// of the BSCC.
std::vector<ValueType> rewardRightSide;
rewardRightSide.reserve(statesNotInBsccs.getNumberOfSetBits());
for (auto state : statesNotInBsccs) {
ValueType reward = zero;
for (auto entry : transitionMatrix.getRow(state)) {
if (statesInBsccs.get(entry.getColumn())) {
reward += entry.getValue() * bsccLra[stateToBsccIndexMap[entry.getColumn()]];
}
}
rewardRightSide.push_back(reward);
}
storm::storage::SparseMatrix<ValueType> rewardEquationSystemMatrix = transitionMatrix.getSubmatrix(false, statesNotInBsccs, statesNotInBsccs, true);
rewardEquationSystemMatrix.convertToEquationSystem();
rewardSolution = std::vector<ValueType>(rewardEquationSystemMatrix.getColumnCount(), one);
{
std::unique_ptr<storm::solver::LinearEquationSolver<ValueType>> solver = linearEquationSolverFactory.create(rewardEquationSystemMatrix);
solver->solveEquationSystem(rewardSolution, rewardRightSide);
}
}
// Fill the result vector.
std::vector<ValueType> result(numOfStates);
auto rewardSolutionIter = rewardSolution.begin();
for (size_t state = 0; state < numOfStates; ++state) {
if (statesInBsccs.get(state)) {
// Assign the value of the bscc the state is in.
result[state] = bsccLra[stateToBsccIndexMap[state]];
} else {
STORM_LOG_ASSERT(rewardSolutionIter != rewardSolution.end(), "Too few elements in solution.");
// Take the value from the reward computation. Since the n-th state not in any bscc is the n-th
// entry in rewardSolution we can just take the next value from the iterator.
result[state] = *rewardSolutionIter;
++rewardSolutionIter;
}
}
return result;
//old implementeation
//now we calculate the long run average for each BSCC in isolation and compute the weighted contribution of the BSCC to the LRA value of all states
//for (uint_fast64_t currentBsccIndex = 0; currentBsccIndex < bsccDecomposition.size(); ++currentBsccIndex) {
// storm::storage::StronglyConnectedComponent const& bscc = bsccDecomposition[currentBsccIndex];
// storm::storage::BitVector statesInThisBscc(numOfStates);
// for (auto const& state : bscc) {
// statesInThisBscc.set(state);
// }
// //ValueType lraForThisBscc = this->computeLraForBSCC(transitionMatrix, psiStates, bscc);
// ValueType lraForThisBscc = bsccLra[currentBsccIndex];
// //the LRA value of a BSCC contributes to the LRA value of a state with the probability of reaching that BSCC from that state
// //thus we add Prob(Eventually statesInThisBscc) * lraForThisBscc to the result vector
// //the reachability probabilities will be zero in other BSCCs, thus we can set the left operand of the until formula to statesNotInBsccs as an optimization
// std::vector<ValueType> reachProbThisBscc = this->computeUntilProbabilitiesHelper(statesNotInBsccs, statesInThisBscc, false);
//
// storm::utility::vector::scaleVectorInPlace<ValueType>(reachProbThisBscc, lraForThisBscc);
// storm::utility::vector::addVectorsInPlace<ValueType>(result, reachProbThisBscc);
//}
//return result;
}
template<typename ValueType>
std::unique_ptr<CheckResult> SparseDtmcPrctlModelChecker<ValueType>::computeLongRunAverage(storm::logic::StateFormula const& stateFormula, bool qualitative, boost::optional<storm::logic::OptimalityType> const& optimalityType) {
std::unique_ptr<CheckResult> subResultPointer = this->check(stateFormula);
ExplicitQualitativeCheckResult const& subResult = subResultPointer->asExplicitQualitativeCheckResult();
return std::unique_ptr<CheckResult>(new ExplicitQuantitativeCheckResult<ValueType>(this->computeLongRunAverageHelper(this->getModel(), this->getModel().getTransitionMatrix(), subResult.getTruthValuesVector(), qualitative, *linearEquationSolverFactory)));
}
template<typename ValueType>
ValueType SparseDtmcPrctlModelChecker<ValueType>::computeLraForBSCC(storm::storage::SparseMatrix<ValueType> const& transitionMatrix, storm::storage::BitVector const& psiStates, storm::storage::StronglyConnectedComponent const& bscc, storm::utility::solver::LinearEquationSolverFactory<ValueType> const& linearEquationSolverFactory) {
//if size is one we can immediately derive the result
if (bscc.size() == 1){
if (psiStates.get(*(bscc.begin()))) {
return storm::utility::one<ValueType>();
} else{
return storm::utility::zero<ValueType>();
}
}
storm::storage::BitVector subsystem = storm::storage::BitVector(transitionMatrix.getRowCount());
subsystem.set(bscc.begin(), bscc.end());
//we now have to solve ((P')^t - I) * x = 0, where P' is the submatrix of transitionMatrix,
// ^t means transose, and I is the identity matrix.
storm::storage::SparseMatrix<ValueType> subsystemMatrix = transitionMatrix.getSubmatrix(false, subsystem, subsystem, true);
subsystemMatrix = subsystemMatrix.transpose();
// subtract 1 on the diagonal and at the same time add a row with all ones to enforce that the result of the equation system is unique
storm::storage::SparseMatrixBuilder<ValueType> equationSystemBuilder(subsystemMatrix.getRowCount() + 1, subsystemMatrix.getColumnCount(), subsystemMatrix.getEntryCount() + subsystemMatrix.getColumnCount());
ValueType one = storm::utility::one<ValueType>();
ValueType zero = storm::utility::zero<ValueType>();
bool foundDiagonalElement = false;
for (uint_fast64_t row = 0; row < subsystemMatrix.getRowCount(); ++row) {
for (auto& entry : subsystemMatrix.getRowGroup(row)) {
if (entry.getColumn() == row) {
equationSystemBuilder.addNextValue(row, entry.getColumn(), entry.getValue() - one);
foundDiagonalElement = true;
} else {
equationSystemBuilder.addNextValue(row, entry.getColumn(), entry.getValue());
}
}
// Throw an exception if a row did not have an element on the diagonal.
STORM_LOG_THROW(foundDiagonalElement, storm::exceptions::InvalidOperationException, "Internal Error, no diagonal entry found.");
}
for (uint_fast64_t column = 0; column + 1 < subsystemMatrix.getColumnCount(); ++column) {
equationSystemBuilder.addNextValue(subsystemMatrix.getRowCount(), column, one);
}
equationSystemBuilder.addNextValue(subsystemMatrix.getRowCount(), subsystemMatrix.getColumnCount() - 1, zero);
subsystemMatrix = equationSystemBuilder.build();
// create x and b for the equation system. setting the last entry of b to one enforces the sum over the unique solution vector is one
std::vector<ValueType> steadyStateDist(subsystemMatrix.getRowCount(), zero);
std::vector<ValueType> b(subsystemMatrix.getRowCount(), zero);
b[subsystemMatrix.getRowCount() - 1] = one;
{
auto solver = linearEquationSolverFactory.create(subsystemMatrix);
solver->solveEquationSystem(steadyStateDist, b);
}
//remove the last entry of the vector as it was just there to enforce the unique solution
steadyStateDist.pop_back();
//calculate the fraction we spend in psi states on the long run
std::vector<ValueType> steadyStateForPsiStates(transitionMatrix.getRowCount() - 1, zero);
storm::utility::vector::setVectorValues(steadyStateForPsiStates, psiStates & subsystem, steadyStateDist);
ValueType result = zero;
for (auto value : steadyStateForPsiStates) {
result += value;
}
return result;
return std::unique_ptr<CheckResult>(new ExplicitQuantitativeCheckResult<ValueType>(SparseCtmcCslModelChecker<ValueType>::computeLongRunAverageHelper(this->getModel(), this->getModel().getTransitionMatrix(), subResult.getTruthValuesVector(), nullptr, qualitative, *linearEquationSolverFactory)));
}
template<typename ValueType>

2
src/modelchecker/prctl/SparseDtmcPrctlModelChecker.h

@ -43,8 +43,6 @@ namespace storm {
static std::vector<ValueType> computeReachabilityRewardsHelper(storm::storage::SparseMatrix<ValueType> const& transitionMatrix, boost::optional<std::vector<ValueType>> const& stateRewardVector, boost::optional<storm::storage::SparseMatrix<ValueType>> const& transitionRewardMatrix, storm::storage::SparseMatrix<ValueType> const& backwardTransitions, storm::storage::BitVector const& targetStates, storm::utility::solver::LinearEquationSolverFactory<ValueType> const& linearEquationSolverFactory, bool qualitative);
static std::vector<ValueType> computeLongRunAverageHelper(storm::models::sparse::DeterministicModel<ValueType> const& model, storm::storage::SparseMatrix<ValueType> const& transitionMatrix, storm::storage::BitVector const& psiStates, bool qualitative, storm::utility::solver::LinearEquationSolverFactory<ValueType> const& linearEquationSolverFactory);
static ValueType computeLraForBSCC(storm::storage::SparseMatrix<ValueType> const& transitionMatrix, storm::storage::BitVector const& goalStates, storm::storage::StronglyConnectedComponent const& bscc, storm::utility::solver::LinearEquationSolverFactory<ValueType> const& linearEquationSolverFactory);
// An object that is used for retrieving linear equation solvers.
std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<ValueType>> linearEquationSolverFactory;
};

36
src/settings/modules/GmmxxEquationSolverSettings.cpp

@ -7,7 +7,7 @@ namespace storm {
namespace modules {
const std::string GmmxxEquationSolverSettings::moduleName = "gmm++";
const std::string GmmxxEquationSolverSettings::techniqueOptionName = "tech";
const std::string GmmxxEquationSolverSettings::techniqueOptionName = "method";
const std::string GmmxxEquationSolverSettings::preconditionOptionName = "precond";
const std::string GmmxxEquationSolverSettings::restartOptionName = "restart";
const std::string GmmxxEquationSolverSettings::maximalIterationsOptionName = "maxiter";
@ -32,38 +32,38 @@ namespace storm {
this->addOption(storm::settings::OptionBuilder(moduleName, absoluteOptionName, false, "Sets whether the relative or the absolute error is considered for detecting convergence.").build());
}
bool GmmxxEquationSolverSettings::isLinearEquationSystemTechniqueSet() const {
bool GmmxxEquationSolverSettings::isLinearEquationSystemMethodSet() const {
return this->getOption(techniqueOptionName).getHasOptionBeenSet();
}
GmmxxEquationSolverSettings::LinearEquationTechnique GmmxxEquationSolverSettings::getLinearEquationSystemTechnique() const {
GmmxxEquationSolverSettings::LinearEquationMethod GmmxxEquationSolverSettings::getLinearEquationSystemMethod() const {
std::string linearEquationSystemTechniqueAsString = this->getOption(techniqueOptionName).getArgumentByName("name").getValueAsString();
if (linearEquationSystemTechniqueAsString == "bicgstab") {
return GmmxxEquationSolverSettings::LinearEquationTechnique::Bicgstab;
return GmmxxEquationSolverSettings::LinearEquationMethod::Bicgstab;
} else if (linearEquationSystemTechniqueAsString == "qmr") {
return GmmxxEquationSolverSettings::LinearEquationTechnique::Qmr;
return GmmxxEquationSolverSettings::LinearEquationMethod::Qmr;
} else if (linearEquationSystemTechniqueAsString == "gmres") {
return GmmxxEquationSolverSettings::LinearEquationTechnique::Gmres;
return GmmxxEquationSolverSettings::LinearEquationMethod::Gmres;
} else if (linearEquationSystemTechniqueAsString == "jacobi") {
return GmmxxEquationSolverSettings::LinearEquationTechnique::Jacobi;
return GmmxxEquationSolverSettings::LinearEquationMethod::Jacobi;
}
STORM_LOG_THROW(false, storm::exceptions::IllegalArgumentValueException, "Unknown solution technique '" << linearEquationSystemTechniqueAsString << "' selected.");
}
bool GmmxxEquationSolverSettings::isPreconditioningTechniqueSet() const {
bool GmmxxEquationSolverSettings::isPreconditioningMethodSet() const {
return this->getOption(preconditionOptionName).getHasOptionBeenSet();
}
GmmxxEquationSolverSettings::PreconditioningTechnique GmmxxEquationSolverSettings::getPreconditioningTechnique() const {
std::string preconditioningTechniqueAsString = this->getOption(preconditionOptionName).getArgumentByName("name").getValueAsString();
if (preconditioningTechniqueAsString == "ilu") {
return GmmxxEquationSolverSettings::PreconditioningTechnique::Ilu;
} else if (preconditioningTechniqueAsString == "diagonal") {
return GmmxxEquationSolverSettings::PreconditioningTechnique::Diagonal;
} else if (preconditioningTechniqueAsString == "none") {
return GmmxxEquationSolverSettings::PreconditioningTechnique::None;
GmmxxEquationSolverSettings::PreconditioningMethod GmmxxEquationSolverSettings::getPreconditioningMethod() const {
std::string PreconditioningMethodAsString = this->getOption(preconditionOptionName).getArgumentByName("name").getValueAsString();
if (PreconditioningMethodAsString == "ilu") {
return GmmxxEquationSolverSettings::PreconditioningMethod::Ilu;
} else if (PreconditioningMethodAsString == "diagonal") {
return GmmxxEquationSolverSettings::PreconditioningMethod::Diagonal;
} else if (PreconditioningMethodAsString == "none") {
return GmmxxEquationSolverSettings::PreconditioningMethod::None;
}
STORM_LOG_THROW(false, storm::exceptions::IllegalArgumentValueException, "Unknown preconditioning technique '" << preconditioningTechniqueAsString << "' selected.");
STORM_LOG_THROW(false, storm::exceptions::IllegalArgumentValueException, "Unknown preconditioning technique '" << PreconditioningMethodAsString << "' selected.");
}
bool GmmxxEquationSolverSettings::isRestartIterationCountSet() const {
@ -99,7 +99,7 @@ namespace storm {
}
bool GmmxxEquationSolverSettings::check() const {
bool optionsSet = isLinearEquationSystemTechniqueSet() || isPreconditioningTechniqueSet() || isRestartIterationCountSet() | isMaximalIterationCountSet() || isPrecisionSet() || isConvergenceCriterionSet();
bool optionsSet = isLinearEquationSystemMethodSet() || isPreconditioningMethodSet() || isRestartIterationCountSet() | isMaximalIterationCountSet() || isPrecisionSet() || isConvergenceCriterionSet();
STORM_LOG_WARN_COND(storm::settings::generalSettings().getEquationSolver() == storm::settings::modules::GeneralSettings::EquationSolver::Gmmxx || !optionsSet, "gmm++ is not selected as the equation solver, so setting options for gmm++ has no effect.");

34
src/settings/modules/GmmxxEquationSolverSettings.h

@ -12,11 +12,11 @@ namespace storm {
*/
class GmmxxEquationSolverSettings : public ModuleSettings {
public:
// An enumeration of all available techniques for solving linear equations.
enum class LinearEquationTechnique { Bicgstab, Qmr, Gmres, Jacobi };
// An enumeration of all available methods for solving linear equations.
enum class LinearEquationMethod { Bicgstab, Qmr, Gmres, Jacobi };
// An enumeration of all available preconditioning techniques.
enum class PreconditioningTechnique { Ilu, Diagonal, None };
// An enumeration of all available preconditioning methods.
enum class PreconditioningMethod { Ilu, Diagonal, None };
// An enumeration of all available convergence criteria.
enum class ConvergenceCriterion { Absolute, Relative };
@ -29,32 +29,32 @@ namespace storm {
GmmxxEquationSolverSettings(storm::settings::SettingsManager& settingsManager);
/*!
* Retrieves whether the linear equation system technique has been set.
* Retrieves whether the linear equation system method has been set.
*
* @return True iff the linear equation system technique has been set.
* @return True iff the linear equation system method has been set.
*/
bool isLinearEquationSystemTechniqueSet() const;
bool isLinearEquationSystemMethodSet() const;
/*!
* Retrieves the technique that is to be used for solving systems of linear equations.
* Retrieves the method that is to be used for solving systems of linear equations.
*
* @return The technique to use.
* @return The method to use.
*/
LinearEquationTechnique getLinearEquationSystemTechnique() const;
LinearEquationMethod getLinearEquationSystemMethod() const;
/*!
* Retrieves whether the preconditioning technique has been set.
* Retrieves whether the preconditioning method has been set.
*
* @return True iff the preconditioning technique has been set.
* @return True iff the preconditioning method has been set.
*/
bool isPreconditioningTechniqueSet() const;
bool isPreconditioningMethodSet() const;
/*!
* Retrieves the technique that is to be used for preconditioning solving systems of linear equations.
* Retrieves the method that is to be used for preconditioning solving systems of linear equations.
*
* @return The technique to use.
* @return The method to use.
*/
PreconditioningTechnique getPreconditioningTechnique() const;
PreconditioningMethod getPreconditioningMethod() const;
/*!
* Retrieves whether the restart iteration count has been set.
@ -64,7 +64,7 @@ namespace storm {
bool isRestartIterationCountSet() const;
/*!
* Retrieves the number of iterations after which restarted techniques are to be restarted.
* Retrieves the number of iterations after which restarted methods are to be restarted.
*
* @return The number of iterations after which to restart.
*/

19
src/settings/modules/NativeEquationSolverSettings.cpp

@ -7,20 +7,23 @@ namespace storm {
namespace modules {
const std::string NativeEquationSolverSettings::moduleName = "native";
const std::string NativeEquationSolverSettings::techniqueOptionName = "tech";
const std::string NativeEquationSolverSettings::techniqueOptionName = "method";
const std::string NativeEquationSolverSettings::omegaOptionName = "soromega";
const std::string NativeEquationSolverSettings::maximalIterationsOptionName = "maxiter";
const std::string NativeEquationSolverSettings::maximalIterationsOptionShortName = "i";
const std::string NativeEquationSolverSettings::precisionOptionName = "precision";
const std::string NativeEquationSolverSettings::absoluteOptionName = "absolute";
NativeEquationSolverSettings::NativeEquationSolverSettings(storm::settings::SettingsManager& settingsManager) : ModuleSettings(settingsManager, moduleName) {
std::vector<std::string> methods = { "jacobi" };
std::vector<std::string> methods = { "jacobi", "gaussseidel", "sor" };
this->addOption(storm::settings::OptionBuilder(moduleName, techniqueOptionName, true, "The method to be used for solving linear equation systems with the native engine. Available are: { jacobi }.").addArgument(storm::settings::ArgumentBuilder::createStringArgument("name", "The name of the method to use.").addValidationFunctionString(storm::settings::ArgumentValidators::stringInListValidator(methods)).setDefaultValueString("jacobi").build()).build());
this->addOption(storm::settings::OptionBuilder(moduleName, maximalIterationsOptionName, false, "The maximal number of iterations to perform before iterative solving is aborted.").setShortName(maximalIterationsOptionShortName).addArgument(storm::settings::ArgumentBuilder::createUnsignedIntegerArgument("count", "The maximal iteration count.").setDefaultValueUnsignedInteger(20000).build()).build());
this->addOption(storm::settings::OptionBuilder(moduleName, precisionOptionName, false, "The precision used for detecting convergence of iterative methods.").addArgument(storm::settings::ArgumentBuilder::createDoubleArgument("value", "The precision to achieve.").setDefaultValueDouble(1e-06).addValidationFunctionDouble(storm::settings::ArgumentValidators::doubleRangeValidatorExcluding(0.0, 1.0)).build()).build());
this->addOption(storm::settings::OptionBuilder(moduleName, omegaOptionName, false, "The omega used for SOR.").addArgument(storm::settings::ArgumentBuilder::createDoubleArgument("value", "The value of the SOR parameter.").setDefaultValueDouble(0.9).addValidationFunctionDouble(storm::settings::ArgumentValidators::doubleRangeValidatorExcluding(0.0, 1.0)).build()).build());
this->addOption(storm::settings::OptionBuilder(moduleName, absoluteOptionName, false, "Sets whether the relative or the absolute error is considered for detecting convergence.").build());
}
@ -28,10 +31,14 @@ namespace storm {
return this->getOption(techniqueOptionName).getHasOptionBeenSet();
}
NativeEquationSolverSettings::LinearEquationTechnique NativeEquationSolverSettings::getLinearEquationSystemTechnique() const {
NativeEquationSolverSettings::LinearEquationMethod NativeEquationSolverSettings::getLinearEquationSystemMethod() const {
std::string linearEquationSystemTechniqueAsString = this->getOption(techniqueOptionName).getArgumentByName("name").getValueAsString();
if (linearEquationSystemTechniqueAsString == "jacobi") {
return NativeEquationSolverSettings::LinearEquationTechnique::Jacobi;
return NativeEquationSolverSettings::LinearEquationMethod::Jacobi;
} else if (linearEquationSystemTechniqueAsString == "gaussseidel") {
return NativeEquationSolverSettings::LinearEquationMethod::GaussSeidel;
} else if (linearEquationSystemTechniqueAsString == "sor") {
return NativeEquationSolverSettings::LinearEquationMethod::SOR;
}
STORM_LOG_THROW(false, storm::exceptions::IllegalArgumentValueException, "Unknown solution technique '" << linearEquationSystemTechniqueAsString << "' selected.");
}
@ -51,6 +58,10 @@ namespace storm {
double NativeEquationSolverSettings::getPrecision() const {
return this->getOption(precisionOptionName).getArgumentByName("value").getValueAsDouble();
}
double NativeEquationSolverSettings::getOmega() const {
return this->getOption(omegaOptionName).getArgumentByName("value").getValueAsDouble();
}
bool NativeEquationSolverSettings::isConvergenceCriterionSet() const {
return this->getOption(absoluteOptionName).getHasOptionBeenSet();

18
src/settings/modules/NativeEquationSolverSettings.h

@ -12,8 +12,8 @@ namespace storm {
*/
class NativeEquationSolverSettings : public ModuleSettings {
public:
// An enumeration of all available techniques for solving linear equations.
enum class LinearEquationTechnique { Jacobi };
// An enumeration of all available methods for solving linear equations.
enum class LinearEquationMethod { Jacobi, GaussSeidel, SOR };
// An enumeration of all available convergence criteria.
enum class ConvergenceCriterion { Absolute, Relative };
@ -33,11 +33,11 @@ namespace storm {
bool isLinearEquationSystemTechniqueSet() const;
/*!
* Retrieves the technique that is to be used for solving systems of linear equations.
* Retrieves the method that is to be used for solving systems of linear equations.
*
* @return The technique to use.
* @return The method to use.
*/
LinearEquationTechnique getLinearEquationSystemTechnique() const;
LinearEquationMethod getLinearEquationSystemMethod() const;
/*!
* Retrieves whether the maximal iteration count has been set.
@ -67,6 +67,13 @@ namespace storm {
*/
double getPrecision() const;
/*!
* Retrieves the value of omega to be used for the SOR method.
*
* @return The value of omega to be used.
*/
double getOmega() const;
/*!
* Retrieves whether the convergence criterion has been set.
*
@ -89,6 +96,7 @@ namespace storm {
private:
// Define the string names of the options as constants.
static const std::string techniqueOptionName;
static const std::string omegaOptionName;
static const std::string maximalIterationsOptionName;
static const std::string maximalIterationsOptionShortName;
static const std::string precisionOptionName;

24
src/solver/GmmxxLinearEquationSolver.cpp

@ -31,24 +31,24 @@ namespace storm {
restart = settings.getRestartIterationCount();
// Determine the method to be used.
storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationTechnique methodAsSetting = settings.getLinearEquationSystemTechnique();
if (methodAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationTechnique::Bicgstab) {
storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationMethod methodAsSetting = settings.getLinearEquationSystemMethod();
if (methodAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationMethod::Bicgstab) {
method = SolutionMethod::Bicgstab;
} else if (methodAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationTechnique::Qmr) {
} else if (methodAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationMethod::Qmr) {
method = SolutionMethod::Qmr;
} else if (methodAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationTechnique::Gmres) {
} else if (methodAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationMethod::Gmres) {
method = SolutionMethod::Gmres;
} else if (methodAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationTechnique::Jacobi) {
} else if (methodAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::LinearEquationMethod::Jacobi) {
method = SolutionMethod::Jacobi;
}
// Check which preconditioner to use.
storm::settings::modules::GmmxxEquationSolverSettings::PreconditioningTechnique preconditionAsSetting = settings.getPreconditioningTechnique();
if (preconditionAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::PreconditioningTechnique::Ilu) {
storm::settings::modules::GmmxxEquationSolverSettings::PreconditioningMethod preconditionAsSetting = settings.getPreconditioningMethod();
if (preconditionAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::PreconditioningMethod::Ilu) {
preconditioner = Preconditioner::Ilu;
} else if (preconditionAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::PreconditioningTechnique::Diagonal) {
} else if (preconditionAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::PreconditioningMethod::Diagonal) {
preconditioner = Preconditioner::Diagonal;
} else if (preconditionAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::PreconditioningTechnique::None) {
} else if (preconditionAsSetting == storm::settings::modules::GmmxxEquationSolverSettings::PreconditioningMethod::None) {
preconditioner = Preconditioner::None;
}
}
@ -149,12 +149,6 @@ namespace storm {
uint_fast64_t GmmxxLinearEquationSolver<ValueType>::solveLinearEquationSystemWithJacobi(storm::storage::SparseMatrix<ValueType> const& A, std::vector<ValueType>& x, std::vector<ValueType> const& b, std::vector<ValueType>* multiplyResult) const {
// Get a Jacobi decomposition of the matrix A.
std::pair<storm::storage::SparseMatrix<ValueType>, std::vector<ValueType>> jacobiDecomposition = A.getJacobiDecomposition();
std::cout << "LU" << std::endl;
std::cout << jacobiDecomposition.first << std::endl;
for (auto const& elem : jacobiDecomposition.second) {
std::cout << elem << std::endl;
}
std::cout << "----" << std::endl;
// Convert the LU matrix to gmm++'s format.
std::unique_ptr<gmm::csr_matrix<ValueType>> gmmLU = storm::adapters::GmmxxAdapter::toGmmxxSparseMatrix<ValueType>(std::move(jacobiDecomposition.first));

131
src/solver/NativeLinearEquationSolver.cpp

@ -15,7 +15,7 @@ namespace storm {
}
template<typename ValueType>
NativeLinearEquationSolver<ValueType>::NativeLinearEquationSolver(storm::storage::SparseMatrix<ValueType> const& A) : A(A) {
NativeLinearEquationSolver<ValueType>::NativeLinearEquationSolver(storm::storage::SparseMatrix<ValueType> const& A, SolutionMethod method) : A(A), method(method) {
// Get the settings object to customize linear solving.
storm::settings::modules::NativeEquationSolverSettings const& settings = storm::settings::nativeEquationSolverSettings();
@ -23,61 +23,94 @@ namespace storm {
maximalNumberOfIterations = settings.getMaximalIterationCount();
precision = settings.getPrecision();
relative = settings.getConvergenceCriterion() == storm::settings::modules::NativeEquationSolverSettings::ConvergenceCriterion::Relative;
// Determine the method to be used.
storm::settings::modules::NativeEquationSolverSettings::LinearEquationTechnique methodAsSetting = settings.getLinearEquationSystemTechnique();
if (methodAsSetting == storm::settings::modules::NativeEquationSolverSettings::LinearEquationTechnique::Jacobi) {
method = SolutionMethod::Jacobi;
}
}
template<typename ValueType>
void NativeLinearEquationSolver<ValueType>::solveEquationSystem(std::vector<ValueType>& x, std::vector<ValueType> const& b, std::vector<ValueType>* multiplyResult) const {
// Get a Jacobi decomposition of the matrix A.
std::pair<storm::storage::SparseMatrix<ValueType>, std::vector<ValueType>> jacobiDecomposition = A.getJacobiDecomposition();
// To avoid copying the contents of the vector in the loop, we create a temporary x to swap with.
bool multiplyResultProvided = true;
std::vector<ValueType>* nextX = multiplyResult;
if (nextX == nullptr) {
nextX = new std::vector<ValueType>(x.size());
multiplyResultProvided = false;
}
std::vector<ValueType> const* copyX = nextX;
std::vector<ValueType>* currentX = &x;
// Target vector for precision calculation.
std::vector<ValueType> tmpX(x.size());
// Set up additional environment variables.
uint_fast64_t iterationCount = 0;
bool converged = false;
while (!converged && iterationCount < maximalNumberOfIterations) {
// Compute D^-1 * (b - LU * x) and store result in nextX.
jacobiDecomposition.first.multiplyWithVector(*currentX, tmpX);
storm::utility::vector::subtractVectors(b, tmpX, tmpX);
storm::utility::vector::multiplyVectorsPointwise(jacobiDecomposition.second, tmpX, *nextX);
if (method == SolutionMethod::SOR || method == SolutionMethod::GaussSeidel) {
// Define the omega used for SOR.
ValueType omega = method == SolutionMethod::SOR ? storm::settings::nativeEquationSolverSettings().getOmega() : storm::utility::one<ValueType>();
// Swap the two pointers as a preparation for the next iteration.
std::swap(nextX, currentX);
// To avoid copying the contents of the vector in the loop, we create a temporary x to swap with.
bool tmpXProvided = true;
std::vector<ValueType>* tmpX = multiplyResult;
if (multiplyResult == nullptr) {
tmpX = new std::vector<ValueType>(x);
tmpXProvided = false;
} else {
*tmpX = x;
}
// Now check if the process already converged within our precision.
converged = storm::utility::vector::equalModuloPrecision<ValueType>(*currentX, *nextX, static_cast<ValueType>(precision), relative);
// Set up additional environment variables.
uint_fast64_t iterationCount = 0;
bool converged = false;
// Increase iteration count so we can abort if convergence is too slow.
++iterationCount;
}
// If the last iteration did not write to the original x we have to swap the contents, because the
// output has to be written to the input parameter x.
if (currentX == copyX) {
std::swap(x, *currentX);
}
// If the vector for the temporary multiplication result was not provided, we need to delete it.
if (!multiplyResultProvided) {
delete copyX;
while (!converged && iterationCount < maximalNumberOfIterations) {
A.performSuccessiveOverRelaxationStep(omega, x, b);
// Now check if the process already converged within our precision.
converged = storm::utility::vector::equalModuloPrecision<ValueType>(x, *tmpX, static_cast<ValueType>(precision), relative);
// If we did not yet converge, we need to copy the contents of x to *tmpX.
if (!converged) {
*tmpX = x;
}
// Increase iteration count so we can abort if convergence is too slow.
++iterationCount;
}
// If the vector for the temporary multiplication result was not provided, we need to delete it.
if (!tmpXProvided) {
delete tmpX;
}
} else {
// Get a Jacobi decomposition of the matrix A.
std::pair<storm::storage::SparseMatrix<ValueType>, std::vector<ValueType>> jacobiDecomposition = A.getJacobiDecomposition();
// To avoid copying the contents of the vector in the loop, we create a temporary x to swap with.
bool multiplyResultProvided = true;
std::vector<ValueType>* nextX = multiplyResult;
if (nextX == nullptr) {
nextX = new std::vector<ValueType>(x.size());
multiplyResultProvided = false;
}
std::vector<ValueType> const* copyX = nextX;
std::vector<ValueType>* currentX = &x;
// Target vector for precision calculation.
std::vector<ValueType> tmpX(x.size());
// Set up additional environment variables.
uint_fast64_t iterationCount = 0;
bool converged = false;
while (!converged && iterationCount < maximalNumberOfIterations) {
// Compute D^-1 * (b - LU * x) and store result in nextX.
jacobiDecomposition.first.multiplyWithVector(*currentX, tmpX);
storm::utility::vector::subtractVectors(b, tmpX, tmpX);
storm::utility::vector::multiplyVectorsPointwise(jacobiDecomposition.second, tmpX, *nextX);
// Swap the two pointers as a preparation for the next iteration.
std::swap(nextX, currentX);
// Now check if the process already converged within our precision.
converged = storm::utility::vector::equalModuloPrecision<ValueType>(*currentX, *nextX, static_cast<ValueType>(precision), relative);
// Increase iteration count so we can abort if convergence is too slow.
++iterationCount;
}
// If the last iteration did not write to the original x we have to swap the contents, because the
// output has to be written to the input parameter x.
if (currentX == copyX) {
std::swap(x, *currentX);
}
// If the vector for the temporary multiplication result was not provided, we need to delete it.
if (!multiplyResultProvided) {
delete copyX;
}
}
}

5
src/solver/NativeLinearEquationSolver.h

@ -14,15 +14,16 @@ namespace storm {
public:
// An enumeration specifying the available solution methods.
enum class SolutionMethod {
Jacobi
Jacobi, GaussSeidel, SOR
};
/*!
* Constructs a linear equation solver with parameters being set according to the settings object.
*
* @param A The matrix defining the coefficients of the linear equation system.
* @param method The method to use for solving linear equations.
*/
NativeLinearEquationSolver(storm::storage::SparseMatrix<ValueType> const& A);
NativeLinearEquationSolver(storm::storage::SparseMatrix<ValueType> const& A, SolutionMethod method = SolutionMethod::Jacobi);
/*!
* Constructs a linear equation solver with the given parameters.

261
src/storage/SparseMatrix.cpp

@ -50,7 +50,7 @@ namespace storm {
void MatrixEntry<IndexType, ValueType>::setValue(ValueType const& value) {
this->entry.second = value;
}
template<typename IndexType, typename ValueType>
std::pair<IndexType, ValueType> const& MatrixEntry<IndexType, ValueType>::getColumnValuePair() const {
return this->entry;
@ -102,7 +102,7 @@ namespace storm {
if (row < lastRow) {
throw storm::exceptions::InvalidArgumentException() << "Illegal call to SparseMatrixBuilder::addNextValue: adding an element in row " << row << ", but an element in row " << lastRow << " has already been added.";
}
// Check that we did not move backwards wrt. to column.
if (row == lastRow && column < lastColumn) {
throw storm::exceptions::InvalidArgumentException() << "Illegal call to SparseMatrixBuilder::addNextValue: adding an element in column " << column << " in row " << row << ", but an element in column " << lastColumn << " has already been added in that row.";
@ -117,14 +117,14 @@ namespace storm {
lastRow = row;
}
lastColumn = column;
// Finally, set the element and increase the current size.
columnsAndValues.emplace_back(column, value);
highestColumn = std::max(highestColumn, column);
++currentEntryCount;
// In case we did not expect this value, we throw an exception.
if (forceInitialDimensions) {
STORM_LOG_THROW(!initialRowCountSet || lastRow < initialRowCount, storm::exceptions::OutOfRangeException, "Cannot insert value at illegal row " << lastRow << ".");
@ -159,14 +159,14 @@ namespace storm {
// as now the indices of row i are always between rowIndications[i] and rowIndications[i + 1], also for
// the first and last row.
rowIndications.push_back(currentEntryCount);
uint_fast64_t columnCount = highestColumn + 1;
if (initialColumnCountSet && forceInitialDimensions) {
STORM_LOG_THROW(columnCount <= initialColumnCount, storm::exceptions::InvalidStateException, "Expected not more than " << initialColumnCount << " columns, but got " << columnCount << ".");
columnCount = std::max(columnCount, initialColumnCount);
}
columnCount = std::max(columnCount, overriddenColumnCount);
uint_fast64_t entryCount = currentEntryCount;
if (initialEntryCountSet && forceInitialDimensions) {
STORM_LOG_THROW(entryCount == initialEntryCount, storm::exceptions::InvalidStateException, "Expected " << initialEntryCount << " entries, but got " << entryCount << ".");
@ -184,12 +184,12 @@ namespace storm {
rowGroupCount = std::max(rowGroupCount, initialRowGroupCount);
}
rowGroupCount = std::max(rowGroupCount, overriddenRowGroupCount);
for (index_type i = currentRowGroup; i <= rowGroupCount; ++i) {
rowGroupIndices.push_back(rowCount);
}
}
return SparseMatrix<ValueType>(columnCount, std::move(rowIndications), std::move(columnsAndValues), std::move(rowGroupIndices), hasCustomRowGrouping);
}
@ -270,14 +270,14 @@ namespace storm {
template<typename ValueType>
SparseMatrix<ValueType>::SparseMatrix(index_type columnCount, std::vector<index_type> const& rowIndications, std::vector<MatrixEntry<index_type, ValueType>> const& columnsAndValues, std::vector<index_type> const& rowGroupIndices, bool nontrivialRowGrouping) : rowCount(rowIndications.size() - 1), columnCount(columnCount), entryCount(columnsAndValues.size()), nonzeroEntryCount(0), columnsAndValues(columnsAndValues), rowIndications(rowIndications), nontrivialRowGrouping(nontrivialRowGrouping), rowGroupIndices(rowGroupIndices) {
this->updateNonzeroEntryCount();
this->updateNonzeroEntryCount();
}
template<typename ValueType>
SparseMatrix<ValueType>::SparseMatrix(index_type columnCount, std::vector<index_type>&& rowIndications, std::vector<MatrixEntry<index_type, ValueType>>&& columnsAndValues, std::vector<index_type>&& rowGroupIndices, bool nontrivialRowGrouping) : rowCount(rowIndications.size() - 1), columnCount(columnCount), entryCount(columnsAndValues.size()), nonzeroEntryCount(0), columnsAndValues(std::move(columnsAndValues)), rowIndications(std::move(rowIndications)), nontrivialRowGrouping(nontrivialRowGrouping), rowGroupIndices(std::move(rowGroupIndices)) {
this->updateNonzeroEntryCount();
this->updateNonzeroEntryCount();
}
template<typename ValueType>
SparseMatrix<ValueType>& SparseMatrix<ValueType>::operator=(SparseMatrix<ValueType> const& other) {
// Only perform assignment if source and target are not the same.
@ -321,7 +321,7 @@ namespace storm {
}
bool equalityResult = true;
equalityResult &= this->getRowCount() == other.getRowCount();
if (!equalityResult) {
return false;
@ -379,35 +379,35 @@ namespace storm {
return entryCount;
}
template<typename T>
uint_fast64_t SparseMatrix<T>::getRowGroupEntryCount(uint_fast64_t const group) const {
uint_fast64_t result = 0;
for (uint_fast64_t row = this->getRowGroupIndices()[group]; row < this->getRowGroupIndices()[group + 1]; ++row) {
result += (this->rowIndications[row + 1] - this->rowIndications[row]);
}
return result;
}
template<typename T>
uint_fast64_t SparseMatrix<T>::getRowGroupEntryCount(uint_fast64_t const group) const {
uint_fast64_t result = 0;
for (uint_fast64_t row = this->getRowGroupIndices()[group]; row < this->getRowGroupIndices()[group + 1]; ++row) {
result += (this->rowIndications[row + 1] - this->rowIndications[row]);
}
return result;
}
template<typename ValueType>
typename SparseMatrix<ValueType>::index_type SparseMatrix<ValueType>::getNonzeroEntryCount() const {
return nonzeroEntryCount;
}
template<typename ValueType>
void SparseMatrix<ValueType>::updateNonzeroEntryCount() const {
//SparseMatrix<ValueType>* self = const_cast<SparseMatrix<ValueType>*>(this);
this->nonzeroEntryCount = 0;
for (auto const& element : *this) {
if (element.getValue() != storm::utility::zero<ValueType>()) {
++this->nonzeroEntryCount;
}
}
}
template<typename ValueType>
void SparseMatrix<ValueType>::updateNonzeroEntryCount(std::make_signed<index_type>::type difference) {
this->nonzeroEntryCount += difference;
}
}
template<typename ValueType>
void SparseMatrix<ValueType>::updateNonzeroEntryCount() const {
//SparseMatrix<ValueType>* self = const_cast<SparseMatrix<ValueType>*>(this);
this->nonzeroEntryCount = 0;
for (auto const& element : *this) {
if (element.getValue() != storm::utility::zero<ValueType>()) {
++this->nonzeroEntryCount;
}
}
}
template<typename ValueType>
void SparseMatrix<ValueType>::updateNonzeroEntryCount(std::make_signed<index_type>::type difference) {
this->nonzeroEntryCount += difference;
}
template<typename ValueType>
typename SparseMatrix<ValueType>::index_type SparseMatrix<ValueType>::getRowGroupCount() const {
@ -423,7 +423,7 @@ namespace storm {
std::vector<typename SparseMatrix<ValueType>::index_type> const& SparseMatrix<ValueType>::getRowGroupIndices() const {
return rowGroupIndices;
}
template<typename ValueType>
void SparseMatrix<ValueType>::makeRowsAbsorbing(storm::storage::BitVector const& rows) {
for (auto row : rows) {
@ -457,7 +457,7 @@ namespace storm {
columnValuePtr->setValue(storm::utility::one<ValueType>());
++columnValuePtr;
for (; columnValuePtr != columnValuePtrEnd; ++columnValuePtr) {
++this->nonzeroEntryCount;
++this->nonzeroEntryCount;
columnValuePtr->setColumn(0);
columnValuePtr->setValue(storm::utility::zero<ValueType>());
}
@ -510,33 +510,42 @@ namespace storm {
return getSubmatrix(rowConstraint, columnConstraint, fakeRowGroupIndices, insertDiagonalElements);
}
}
template<typename ValueType>
SparseMatrix<ValueType> SparseMatrix<ValueType>::getSubmatrix(storm::storage::BitVector const& rowGroupConstraint, storm::storage::BitVector const& columnConstraint, std::vector<index_type> const& rowGroupIndices, bool insertDiagonalEntries) const {
uint_fast64_t submatrixColumnCount = columnConstraint.getNumberOfSetBits();
// Start by creating a temporary vector that stores for each index whose bit is set to true the number of
// bits that were set before that particular index.
std::vector<index_type> bitsSetBeforeIndex;
bitsSetBeforeIndex.reserve(columnCount);
std::vector<index_type> columnBitsSetBeforeIndex;
columnBitsSetBeforeIndex.reserve(columnCount);
// Compute the information to fill this vector.
index_type lastIndex = 0;
index_type currentNumberOfSetBits = 0;
// If we are requested to add missing diagonal entries, we need to make sure the corresponding rows are also
// taken.
// storm::storage::BitVector columnBitCountConstraint = columnConstraint;
// if (insertDiagonalEntries) {
// columnBitCountConstraint |= rowGroupConstraint;
// }
for (auto index : columnConstraint) {
while (lastIndex <= index) {
bitsSetBeforeIndex.push_back(currentNumberOfSetBits);
columnBitsSetBeforeIndex.push_back(currentNumberOfSetBits);
++lastIndex;
}
++currentNumberOfSetBits;
}
std::vector<index_type>* rowBitsSetBeforeIndex;
if (&rowGroupConstraint == &columnConstraint) {
rowBitsSetBeforeIndex = &columnBitsSetBeforeIndex;
} else {
rowBitsSetBeforeIndex = new std::vector<index_type>(rowCount);
lastIndex = 0;
currentNumberOfSetBits = 0;
for (auto index : rowGroupConstraint) {
while (lastIndex <= index) {
rowBitsSetBeforeIndex->push_back(currentNumberOfSetBits);
++lastIndex;
}
++currentNumberOfSetBits;
}
}
// Then, we need to determine the number of entries and the number of rows of the submatrix.
index_type subEntries = 0;
@ -551,7 +560,7 @@ namespace storm {
if (columnConstraint.get(it->getColumn())) {
++subEntries;
if (index == it->getColumn()) {
if (columnBitsSetBeforeIndex[it->getColumn()] == (*rowBitsSetBeforeIndex)[index]) {
foundDiagonalElement = true;
}
}
@ -580,13 +589,13 @@ namespace storm {
for (const_iterator it = this->begin(i), ite = this->end(i); it != ite; ++it) {
if (columnConstraint.get(it->getColumn())) {
if (index == it->getColumn()) {
if (columnBitsSetBeforeIndex[it->getColumn()] == (*rowBitsSetBeforeIndex)[index]) {
insertedDiagonalElement = true;
} else if (insertDiagonalEntries && !insertedDiagonalElement && it->getColumn() > index) {
matrixBuilder.addNextValue(rowCount, rowCount, storm::utility::zero<ValueType>());
} else if (insertDiagonalEntries && !insertedDiagonalElement && columnBitsSetBeforeIndex[it->getColumn()] > (*rowBitsSetBeforeIndex)[index]) {
matrixBuilder.addNextValue(rowCount, rowGroupCount, storm::utility::zero<ValueType>());
insertedDiagonalElement = true;
}
matrixBuilder.addNextValue(rowCount, bitsSetBeforeIndex[it->getColumn()], it->getValue());
matrixBuilder.addNextValue(rowCount, columnBitsSetBeforeIndex[it->getColumn()], it->getValue());
}
}
if (insertDiagonalEntries && !insertedDiagonalElement && rowGroupCount < submatrixColumnCount) {
@ -597,6 +606,11 @@ namespace storm {
++rowGroupCount;
}
// If the constraints were not the same, we have allocated some additional memory we need to free now.
if (&rowGroupConstraint != &columnConstraint) {
delete rowBitsSetBeforeIndex;
}
return matrixBuilder.build();
}
@ -655,12 +669,12 @@ namespace storm {
SparseMatrix<ValueType> SparseMatrix<ValueType>::transpose(bool joinGroups, bool keepZeros) const {
index_type rowCount = this->getColumnCount();
index_type columnCount = joinGroups ? this->getRowGroupCount() : this->getRowCount();
if (keepZeros) {
index_type entryCount = this->getEntryCount();
} else {
this->updateNonzeroEntryCount();
index_type entryCount = this->getNonzeroEntryCount();
}
if (keepZeros) {
index_type entryCount = this->getEntryCount();
} else {
this->updateNonzeroEntryCount();
index_type entryCount = this->getNonzeroEntryCount();
}
std::vector<index_type> rowIndications(rowCount + 1);
std::vector<MatrixEntry<index_type, ValueType>> columnsAndValues(entryCount);
@ -668,7 +682,7 @@ namespace storm {
// First, we need to count how many entries each column has.
for (index_type group = 0; group < columnCount; ++group) {
for (auto const& transition : joinGroups ? this->getRowGroup(group) : this->getRow(group)) {
if (transition.getValue() != storm::utility::zero<ValueType>() || keepZeros) {
if (transition.getValue() != storm::utility::zero<ValueType>() || keepZeros) {
++rowIndications[transition.getColumn() + 1];
}
}
@ -687,7 +701,7 @@ namespace storm {
// Now we are ready to actually fill in the values of the transposed matrix.
for (index_type group = 0; group < columnCount; ++group) {
for (auto const& transition : joinGroups ? this->getRowGroup(group) : this->getRow(group)) {
if (transition.getValue() != storm::utility::zero<ValueType>() || keepZeros) {
if (transition.getValue() != storm::utility::zero<ValueType>() || keepZeros) {
columnsAndValues[nextIndices[transition.getColumn()]] = std::make_pair(group, transition.getValue());
nextIndices[transition.getColumn()]++;
}
@ -700,10 +714,10 @@ namespace storm {
}
storm::storage::SparseMatrix<ValueType> transposedMatrix(columnCount, std::move(rowIndications), std::move(columnsAndValues), std::move(rowGroupIndices), false);
return transposedMatrix;
}
template<typename ValueType>
void SparseMatrix<ValueType>::convertToEquationSystem() {
invertDiagonal();
@ -713,22 +727,22 @@ namespace storm {
template<typename ValueType>
void SparseMatrix<ValueType>::invertDiagonal() {
// Now iterate over all row groups and set the diagonal elements to the inverted value.
// If there is a row without the diagonal element, an exception is thrown.
ValueType one = storm::utility::one<ValueType>();
ValueType zero = storm::utility::zero<ValueType>();
// If there is a row without the diagonal element, an exception is thrown.
ValueType one = storm::utility::one<ValueType>();
ValueType zero = storm::utility::zero<ValueType>();
bool foundDiagonalElement = false;
for (index_type group = 0; group < this->getRowGroupCount(); ++group) {
for (auto& entry : this->getRowGroup(group)) {
if (entry.getColumn() == group) {
if (entry.getValue() == one) {
--this->nonzeroEntryCount;
entry.setValue(zero);
} else if (entry.getValue() == zero) {
++this->nonzeroEntryCount;
entry.setValue(one);
} else {
entry.setValue(one - entry.getValue());
}
if (entry.getValue() == one) {
--this->nonzeroEntryCount;
entry.setValue(zero);
} else if (entry.getValue() == zero) {
++this->nonzeroEntryCount;
entry.setValue(one);
} else {
entry.setValue(one - entry.getValue());
}
foundDiagonalElement = true;
}
}
@ -758,7 +772,7 @@ namespace storm {
for (index_type group = 0; group < this->getRowGroupCount(); ++group) {
for (auto& entry : this->getRowGroup(group)) {
if (entry.getColumn() == group) {
--this->nonzeroEntryCount;
--this->nonzeroEntryCount;
entry.setValue(storm::utility::zero<ValueType>());
}
}
@ -839,15 +853,28 @@ namespace storm {
typename std::vector<ValueType>::iterator resultIterator = result.begin();
typename std::vector<ValueType>::iterator resultIteratorEnd = result.end();
for (; resultIterator != resultIteratorEnd; ++rowIterator, ++resultIterator) {
*resultIterator = storm::utility::zero<ValueType>();
for (ite = this->begin() + *(rowIterator + 1); it != ite; ++it) {
*resultIterator += it->getValue() * vector[it->getColumn()];
// If the vector to multiply with and the target vector are actually the same, we need an auxiliary variable
// to store the intermediate result.
if (&vector == &result) {
for (; resultIterator != resultIteratorEnd; ++rowIterator, ++resultIterator) {
ValueType tmpValue = storm::utility::zero<ValueType>();
for (ite = this->begin() + *(rowIterator + 1); it != ite; ++it) {
tmpValue += it->getValue() * vector[it->getColumn()];
}
*resultIterator = tmpValue;
}
} else {
for (; resultIterator != resultIteratorEnd; ++rowIterator, ++resultIterator) {
*resultIterator = storm::utility::zero<ValueType>();
for (ite = this->begin() + *(rowIterator + 1); it != ite; ++it) {
*resultIterator += it->getValue() * vector[it->getColumn()];
}
}
}
}
#ifdef STORM_HAVE_INTELTBB
template<typename ValueType>
void SparseMatrix<ValueType>::multiplyWithVectorParallel(std::vector<ValueType> const& vector, std::vector<ValueType>& result) const {
@ -872,7 +899,37 @@ namespace storm {
});
}
#endif
template<typename ValueType>
void SparseMatrix<ValueType>::performSuccessiveOverRelaxationStep(ValueType omega, std::vector<ValueType>& x, std::vector<ValueType> const& b) const {
const_iterator it = this->begin();
const_iterator ite;
std::vector<index_type>::const_iterator rowIterator = rowIndications.begin();
typename std::vector<ValueType>::const_iterator bIt = b.begin();
typename std::vector<ValueType>::const_iterator bIte = b.end();
typename std::vector<ValueType>::iterator resultIterator = x.begin();
typename std::vector<ValueType>::iterator resultIteratorEnd = x.end();
// If the vector to multiply with and the target vector are actually the same, we need an auxiliary variable
// to store the intermediate result.
index_type currentRow = 0;
for (; resultIterator != resultIteratorEnd; ++rowIterator, ++resultIterator, ++bIt) {
ValueType tmpValue = storm::utility::zero<ValueType>();
ValueType diagonalElement = storm::utility::zero<ValueType>();
for (ite = this->begin() + *(rowIterator + 1); it != ite; ++it) {
if (it->getColumn() != currentRow) {
tmpValue += it->getValue() * x[it->getColumn()];
} else {
diagonalElement += it->getValue();
}
}
*resultIterator = ((storm::utility::one<ValueType>() - omega) * *resultIterator) + (omega / diagonalElement) * (*bIt - tmpValue);
++currentRow;
}
}
template<typename ValueType>
void SparseMatrix<ValueType>::multiplyVectorWithMatrix(std::vector<value_type> const& vector, std::vector<value_type>& result) const {
const_iterator it = this->begin();
@ -982,7 +1039,7 @@ namespace storm {
if (this->getRowCount() != matrix.getRowCount()) return false;
if (this->getColumnCount() != matrix.getColumnCount()) return false;
if (this->getRowGroupIndices() != matrix.getRowGroupIndices()) return false;
// Check the subset property for all rows individually.
for (index_type row = 0; row < this->getRowCount(); ++row) {
for (const_iterator it1 = this->begin(row), ite1 = this->end(row), it2 = matrix.begin(row), ite2 = matrix.end(row); it1 != ite1; ++it1) {
@ -1054,21 +1111,21 @@ namespace storm {
}
// Explicitly instantiate the entry, builder and the matrix.
//double
template class MatrixEntry<typename SparseMatrix<double>::index_type, double>;
template std::ostream& operator<<(std::ostream& out, MatrixEntry<uint_fast64_t, double> const& entry);
template class SparseMatrixBuilder<double>;
template class SparseMatrix<double>;
template std::ostream& operator<<(std::ostream& out, SparseMatrix<double> const& matrix);
//float
template class MatrixEntry<typename SparseMatrix<float>::index_type, float>;
template std::ostream& operator<<(std::ostream& out, MatrixEntry<uint_fast64_t, float> const& entry);
template class SparseMatrixBuilder<float>;
template class SparseMatrix<float>;
template std::ostream& operator<<(std::ostream& out, SparseMatrix<float> const& matrix);
//int
template class MatrixEntry<typename SparseMatrix<int>::index_type, int>;
template std::ostream& operator<<(std::ostream& out, MatrixEntry<uint_fast64_t, int> const& entry);
//double
template class MatrixEntry<typename SparseMatrix<double>::index_type, double>;
template std::ostream& operator<<(std::ostream& out, MatrixEntry<uint_fast64_t, double> const& entry);
template class SparseMatrixBuilder<double>;
template class SparseMatrix<double>;
template std::ostream& operator<<(std::ostream& out, SparseMatrix<double> const& matrix);
//float
template class MatrixEntry<typename SparseMatrix<float>::index_type, float>;
template std::ostream& operator<<(std::ostream& out, MatrixEntry<uint_fast64_t, float> const& entry);
template class SparseMatrixBuilder<float>;
template class SparseMatrix<float>;
template std::ostream& operator<<(std::ostream& out, SparseMatrix<float> const& matrix);
//int
template class MatrixEntry<typename SparseMatrix<int>::index_type, int>;
template std::ostream& operator<<(std::ostream& out, MatrixEntry<uint_fast64_t, int> const& entry);
template class SparseMatrixBuilder<int>;
template class SparseMatrix<int>;
template std::ostream& operator<<(std::ostream& out, SparseMatrix<int> const& matrix);

9
src/storage/SparseMatrix.h

@ -692,6 +692,15 @@ namespace storm {
*/
void multiplyVectorWithMatrix(std::vector<value_type> const& vector, std::vector<value_type>& result) const;
/*!
* Performs one step of the successive over-relaxation technique.
*
* @param omega The Omega parameter for SOR.
* @param x The current solution vector. The result will be written to the very same vector.
* @param b The 'right-hand side' of the problem.
*/
void performSuccessiveOverRelaxationStep(ValueType omega, std::vector<ValueType>& x, std::vector<ValueType> const& b) const;
/*!
* Multiplies the matrix with the given vector in a sequential way and writes the result to the given result
* vector.

21
src/utility/solver.cpp

@ -41,9 +41,28 @@ namespace storm {
return std::unique_ptr<storm::solver::LinearEquationSolver<ValueType>>(new storm::solver::GmmxxLinearEquationSolver<ValueType>(matrix));
}
template<typename ValueType>
NativeLinearEquationSolverFactory<ValueType>::NativeLinearEquationSolverFactory() {
switch (storm::settings::nativeEquationSolverSettings().getLinearEquationSystemMethod()) {
case settings::modules::NativeEquationSolverSettings::LinearEquationMethod::Jacobi:
this->method = storm::solver::NativeLinearEquationSolver<ValueType>::SolutionMethod::Jacobi;
break;
case settings::modules::NativeEquationSolverSettings::LinearEquationMethod::GaussSeidel:
this->method = storm::solver::NativeLinearEquationSolver<ValueType>::SolutionMethod::GaussSeidel;
case settings::modules::NativeEquationSolverSettings::LinearEquationMethod::SOR:
this->method = storm::solver::NativeLinearEquationSolver<ValueType>::SolutionMethod::SOR;
}
omega = storm::settings::nativeEquationSolverSettings().getOmega();
}
template<typename ValueType>
NativeLinearEquationSolverFactory<ValueType>::NativeLinearEquationSolverFactory(typename storm::solver::NativeLinearEquationSolver<ValueType>::SolutionMethod method, ValueType omega) : method(method), omega(omega) {
// Intentionally left empty.
}
template<typename ValueType>
std::unique_ptr<storm::solver::LinearEquationSolver<ValueType>> NativeLinearEquationSolverFactory<ValueType>::create(storm::storage::SparseMatrix<ValueType> const& matrix) const {
return std::unique_ptr<storm::solver::LinearEquationSolver<ValueType>>(new storm::solver::NativeLinearEquationSolver<ValueType>(matrix));
return std::unique_ptr<storm::solver::LinearEquationSolver<ValueType>>(new storm::solver::NativeLinearEquationSolver<ValueType>(matrix, method));
}
template<typename ValueType>

9
src/utility/solver.h

@ -5,10 +5,12 @@
#include "src/solver/SymbolicLinearEquationSolver.h"
#include "src/solver/LinearEquationSolver.h"
#include "src/solver/NativeLinearEquationSolver.h"
#include "src/solver/MinMaxLinearEquationSolver.h"
#include "src/solver/LpSolver.h"
#include "src/storage/dd/DdType.h"
#include "src/settings/modules/NativeEquationSolverSettings.h"
#include "src/exceptions/InvalidSettingsException.h"
@ -42,7 +44,14 @@ namespace storm {
template<typename ValueType>
class NativeLinearEquationSolverFactory : public LinearEquationSolverFactory<ValueType> {
public:
NativeLinearEquationSolverFactory();
NativeLinearEquationSolverFactory(typename storm::solver::NativeLinearEquationSolver<ValueType>::SolutionMethod method, ValueType omega);
virtual std::unique_ptr<storm::solver::LinearEquationSolver<ValueType>> create(storm::storage::SparseMatrix<ValueType> const& matrix) const override;
private:
typename storm::solver::NativeLinearEquationSolver<ValueType>::SolutionMethod method;
ValueType omega;
};
template<typename ValueType>

9
test/functional/modelchecker/NativeDtmcPrctlModelCheckerTest.cpp

@ -1,6 +1,7 @@
#include "gtest/gtest.h"
#include "storm-config.h"
#include "src/settings/SettingMemento.h"
#include "src/logic/Formulas.h"
#include "src/utility/solver.h"
#include "src/modelchecker/prctl/SparseDtmcPrctlModelChecker.h"
@ -144,7 +145,7 @@ TEST(SparseDtmcPrctlModelCheckerTest, LRASingleBscc) {
dtmc.reset(new storm::models::sparse::Dtmc<double>(transitionMatrix, ap, boost::none, boost::none, boost::none));
storm::modelchecker::SparseDtmcPrctlModelChecker<double> checker(*dtmc, std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<double>>(new storm::utility::solver::NativeLinearEquationSolverFactory<double>()));
storm::modelchecker::SparseDtmcPrctlModelChecker<double> checker(*dtmc, std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<double>>(new storm::utility::solver::NativeLinearEquationSolverFactory<double>(storm::solver::NativeLinearEquationSolver<double>::SolutionMethod::SOR, 0.9)));
auto labelFormula = std::make_shared<storm::logic::AtomicLabelFormula>("a");
auto lraFormula = std::make_shared<storm::logic::LongRunAverageOperatorFormula>(labelFormula);
@ -169,7 +170,7 @@ TEST(SparseDtmcPrctlModelCheckerTest, LRASingleBscc) {
dtmc.reset(new storm::models::sparse::Dtmc<double>(transitionMatrix, ap, boost::none, boost::none, boost::none));
storm::modelchecker::SparseDtmcPrctlModelChecker<double> checker(*dtmc, std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<double>>(new storm::utility::solver::NativeLinearEquationSolverFactory<double>()));
storm::modelchecker::SparseDtmcPrctlModelChecker<double> checker(*dtmc, std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<double>>(new storm::utility::solver::NativeLinearEquationSolverFactory<double>(storm::solver::NativeLinearEquationSolver<double>::SolutionMethod::SOR, 0.9)));
auto labelFormula = std::make_shared<storm::logic::AtomicLabelFormula>("a");
auto lraFormula = std::make_shared<storm::logic::LongRunAverageOperatorFormula>(labelFormula);
@ -194,7 +195,7 @@ TEST(SparseDtmcPrctlModelCheckerTest, LRASingleBscc) {
dtmc.reset(new storm::models::sparse::Dtmc<double>(transitionMatrix, ap, boost::none, boost::none, boost::none));
storm::modelchecker::SparseDtmcPrctlModelChecker<double> checker(*dtmc, std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<double>>(new storm::utility::solver::NativeLinearEquationSolverFactory<double>()));
storm::modelchecker::SparseDtmcPrctlModelChecker<double> checker(*dtmc, std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<double>>(new storm::utility::solver::NativeLinearEquationSolverFactory<double>(storm::solver::NativeLinearEquationSolver<double>::SolutionMethod::SOR, 0.9)));
auto labelFormula = std::make_shared<storm::logic::AtomicLabelFormula>("a");
auto lraFormula = std::make_shared<storm::logic::LongRunAverageOperatorFormula>(labelFormula);
@ -255,7 +256,7 @@ TEST(SparseDtmcPrctlModelCheckerTest, LRA) {
mdp.reset(new storm::models::sparse::Dtmc<double>(transitionMatrix, ap, boost::none, boost::none, boost::none));
storm::modelchecker::SparseDtmcPrctlModelChecker<double> checker(*mdp, std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<double>>(new storm::utility::solver::NativeLinearEquationSolverFactory<double>()));
storm::modelchecker::SparseDtmcPrctlModelChecker<double> checker(*mdp, std::unique_ptr<storm::utility::solver::LinearEquationSolverFactory<double>>(new storm::utility::solver::NativeLinearEquationSolverFactory<double>(storm::solver::NativeLinearEquationSolver<double>::SolutionMethod::SOR, 0.9)));
auto labelFormula = std::make_shared<storm::logic::AtomicLabelFormula>("a");
auto lraFormula = std::make_shared<storm::logic::LongRunAverageOperatorFormula>(labelFormula);

Loading…
Cancel
Save