d3ece2a2e5
Better simplification of prism commands.
5 years ago
0e91887ebb
Queried the termination flag in a few more places.
5 years ago
4585f8f555
One more fix for AcyclicSolverHelper.
5 years ago
7766a96783
Fixes for Acylic equation solvers.
5 years ago
bbc6f8b786
Fixed invalid memory access when applying BitVector::resize on BitVectors of length 0.
5 years ago
99f2344da9
Use acyclic solver in various Markov automata methods.
5 years ago
c83721066c
Use acyclic solver in reward bounded properties.
5 years ago
53db0b1f22
Added AcyclicMinMaxLinearEquationSolver and AcyclicLinearEquationSolver which are optimized for many calls on an acyclic model.
5 years ago
31cbe14d3c
Multiplier: Added a flag to specify whether gaussSeidel style multiplications should be performed forward or backwards.
5 years ago
d88e7e9951
Explicit header files to include all defined environments
5 years ago
14f07a2d1a
Unif+: Update kappa only based on the results at the initial state
6 years ago
dd958bcedd
Changed default of the unifpluskappa
6 years ago
c399c31c52
Added missing include
6 years ago
6f62e8d402
Support abort in model building
6 years ago
e65e5587f0
Support for abort in Gmm++ by throwing exception
6 years ago
823ae23180
Use updateStatus in more cases
6 years ago
f50a7a424b
General updateStatus function in AbstractEquationSolver
6 years ago
c6b984ca51
Do not perform the conversion from a prism program to a jani model twice.
6 years ago
b745b10b77
Moved reportStatus() and updateStatusIfNotConverged() to AbstractEquationSolver
6 years ago
7e9029e5bd
Optimization for PRISM model building: Avoid evaluating unnecessarily many guards.
6 years ago
4012094f9d
Use new time bounded environment also in the ctmc solver.
6 years ago
3debbbc64d
Added more abortion checks
6 years ago
d35d31ce94
Added SolverStatus::Abort for abort signal handling
6 years ago
acd5a94162
Use general SolverStatus in StandardGameSolver
6 years ago
8168b9d125
Using OVI as default
6 years ago
0dd1c70e12
Set waiting time after signal with flag --signal-timeout
6 years ago
383e2172d4
Added OVI for linear equation systems (i.e. DTMC/CTMC)
6 years ago
7c49edb68f
Put most of the optimistic value iteration code into a new helper file
6 years ago
6c095e757a
Fixed problem with Windows linebreak \r\n, because this is still a problem in 2020
6 years ago
502b637df4
Ovi: Use an additional auxiliary row group vector (to allow caching).
6 years ago
ee82f30550
Removed Debug output
6 years ago
0ba0d94b84
Small update for the default settings.
6 years ago
33975c181e
Fixes and improvements in the new unif+ implementation.
6 years ago
52857d2701
FoxGlyn: Print an error message instead of throwing an exception in cases where an underflow is possible.
6 years ago
e54a035ab9
SolverEnvironment: Added the switch `forceExact` to switch to exact solving methods more conveniently.
6 years ago
c3184d3afa
Consider relevant states in unif+
6 years ago
ca59635f8a
Using new time bounded environment also for IMCA method.
6 years ago
03bf55f4ab
Bugfix in new unif+ implementation
6 years ago
f8fbf7ace4
Refactored unif+ implementation (introduced relative precision criterion)
6 years ago
bb94110b74
MarkovAutomaton model checkers: Enable consideration of psiStates.
6 years ago
60670e1fb2
SparseMatrix: fixed getConstraintRowSumVector which did not allocate enough space before filling the resulting vector.
6 years ago
f0d1aff610
Added new settings and environment module for time-bounded settings.
6 years ago
ce298fa782
Moved signal handling to own file to prevent problems with global static variables being non-unique
6 years ago
05471d94fd
Optimizations in JaniNextStateGenerator that avoid unnecessary (re-) allocations of memory. Moreover, before collecting the complete set of enabled edges for each automaton, we first check whether each automaton (that synchronizes with that action) has at least one enabled edge. This avoids checking unnecessarily many edge guards.
6 years ago
e79035c71b
Enabled short circuit optimization for & (and) and | (or) in ExprtkExpressionEvaluator
6 years ago
d92e6b97e5
updated exprtk
6 years ago
e945f28a86
Using raw pointers for manager in Variable.h since weak_ptr::lock() often seems to be a bottle neck during, e.g., model building.
6 years ago
5d8419336f
InternalAdds: Added a comment related to GitHub issue #64
6 years ago
36c410875c
Revert "InternalAdds: Making the different splitIntoGroups implementations more consistent to each other (in the sense that the Dd is traversed in the same order)."
This reverts commit cefe43f2bf#64
6 years ago
1f98f6c557
Reverted 'optimization' for Prob1Max (since that didn't work out).
6 years ago
Tim Quatmann
Matthias Volk