13template<
typename PomdpModelType,
typename BeliefValueType =
typename PomdpModelType::ValueType,
14 typename BeliefMdpValueType =
typename PomdpModelType::ValueType>
48 uint64_t resolution,
bool useDynamic,
76 PomdpModelType
const& inputPomdp;
BeliefBasedModelChecker(PomdpModelType const &pomdp)
Creates a checker for a canonic POMDP.
BeliefBasedModelCheckerResult< BeliefMdpValueType > checkRewardAwareDiscretize(storm::Environment const &env, PropertyInformation const &propertyInformation, storm::pomdp::beliefs::BeliefBasedModelCheckerOptions< BeliefMdpValueType > const &options, uint64_t resolution, bool useDynamic, storage::BeliefExplorationBounds< typename PomdpModelType::ValueType > const &valueBounds, std::vector< std::string > const &relevantRewardModelNames)
Combines reward-aware exploration with Freudenthal triangulation discretization.
typename PomdpModelType::ValueType PomdpValueType
BeliefBasedModelCheckerResult< BeliefMdpValueType > checkDiscretize(storm::Environment const &env, PropertyInformation const &propertyInformation, storm::pomdp::beliefs::BeliefBasedModelCheckerOptions< BeliefMdpValueType > const &options, uint64_t resolution, bool useDynamic, storage::BeliefExplorationBounds< PomdpValueType > const &valueBounds)
Explores the belief space and discretises beliefs using the Freudenthal triangualtion approximation.
BeliefBasedModelCheckerResult< BeliefMdpValueType > checkUnfold(storm::Environment const &env, PropertyInformation const &propertyInformation, BeliefBasedModelCheckerOptions< BeliefMdpValueType > const &options, storage::BeliefExplorationBounds< PomdpValueType > const &valueBounds)
Explores the belief MDP by unfolding the belief space.
BeliefBasedModelCheckerResult< BeliefMdpValueType > checkRewardAwareUnfold(storm::Environment const &env, PropertyInformation const &propertyInformation, storm::pomdp::beliefs::BeliefBasedModelCheckerOptions< BeliefMdpValueType > const &options, storage::BeliefExplorationBounds< typename PomdpModelType::ValueType > const &valueBounds, std::vector< std::string > const &relevantRewardModelNames)
Explores a reward-aware belief MDP, splitting beliefs before successor generation by their reward vec...
Result of a belief-based model-checking run.
Struct for storing precomputed values bounding the actual values on the POMDP.