Tim Quatmann
|
53db0b1f22
|
Added AcyclicMinMaxLinearEquationSolver and AcyclicLinearEquationSolver which are optimized for many calls on an acyclic model.
|
5 years ago |
Jan Erik Karuc
|
4fdfc37341
|
Factory, Testing Environment (Topological Excluded)
|
5 years ago |
Tim Quatmann
|
bc623d1203
|
MinMaxLinearEquationSolver: Added a flag 'hasNoEndComponent' that is true if the system is known to have no end components. This decides if policy iteration does require a valid initial scheduler.
Renamed the 'hasNoEndComponents' solver requirement to 'hasUniqueSolution' as this is the actual thing we require for, e.g. sound value iteration.
|
6 years ago |
TimQu
|
6e3639c8f1
|
Added new minmax method: Vi-to-Pi, which first performs value iteration with doubles, to find a good initial policy for (potentially exact) policy iteration.
|
6 years ago |
TimQu
|
87fa9908bf
|
Fixed an issue where scheduler generation in MDPs was not possible due to end components even if there actually were no end components.
|
7 years ago |
TimQu
|
64ba34a397
|
removed multiplication support from minmax equation solvers. Also removed Factories.
|
7 years ago |
TimQu
|
ea25c8fd2e
|
renamed 'sound value iteration' to 'interval iteration' and 'quick sound value iteration' to 'sound value iteration'
|
7 years ago |
TimQu
|
ef893abca6
|
fixed a case for topologicalMinMaxSolver with exact arithmetic
|
7 years ago |
TimQu
|
a2bd1e0026
|
renamed argument from getRequirements so that it is easier to understand
|
7 years ago |
TimQu
|
2d910b79ed
|
Introduced new topological min max solver
|
7 years ago |
TimQu
|
4ab47671f5
|
Renamed TopologicalMinMaxLinearEquationSolver -> TopologicalCudaMinMaxLinearEquationSolver
|
7 years ago |
TimQu
|
9c96bd0a1c
|
First implementation of quick value iteration for MinMax Equation systems
|
7 years ago |
TimQu
|
776ce4c8bb
|
Checking requirements of a linear equation solver now depends on whether we want to do multiplication or equation solving. This was necessary to get the correct requirements of a MinMaxSolver that only uses the underlying linear equation solver for multiplication.
|
8 years ago |
TimQu
|
bb63ac6089
|
Linear equation solver + game solvers now respect the environment as well
|
8 years ago |
TimQu
|
42cea9c688
|
better subenvironments
|
8 years ago |
TimQu
|
6d23c79737
|
Making libstorm compile again
|
8 years ago |
TimQu
|
fd8c99b989
|
Introducing Environment in MinMaxSolvers and ModelCheckers
|
8 years ago |
TimQu
|
17d6835477
|
removed acyclic minmax method as it is not much better then gauss-seidel style multiplications
|
8 years ago |
TimQu
|
33585c811f
|
MinMax Solver requirements now respect whether the solution is known to be unique or not.
|
8 years ago |
dehnert
|
e557a8e069
|
started on EC elimination for hybrid engine
|
8 years ago |
dehnert
|
afd2acd06b
|
more work on Kwek-Mehlhorn approach
|
8 years ago |
dehnert
|
b3f0aa511e
|
started on Kwek-Mehlhorn-based exact value computation
|
8 years ago |
dehnert
|
2d41de479e
|
added progress outputs to iterative solvers
|
8 years ago |
dehnert
|
52d729b1c7
|
upper bounds computation for reachability rewards in sparse MDPs
|
8 years ago |
dehnert
|
d25cc4b05f
|
first version of sound value iteration
|
8 years ago |
dehnert
|
9d95d2adcf
|
first version of multiply-and-reduce (only for native)
|
8 years ago |
dehnert
|
4c5cdfeafc
|
Sparse MDP helper now also respects solver requirements for reachability rewards
|
8 years ago |
dehnert
|
74eeaa7f81
|
computing unbounded until on MDPs with the sparse helper now respects solver requirements
|
8 years ago |
dehnert
|
569b0122b8
|
introduced different minmax equation system types for requirement retrieval
|
8 years ago |
dehnert
|
4adee85fa5
|
added checking requirements of MinMax solvers to model checker helpers
|
8 years ago |
dehnert
|
3829b58e0d
|
introduced top-level solve equations function to centrally check for requirements
|
8 years ago |
dehnert
|
72234e96b2
|
started on requirements for MinMax solvers
|
8 years ago |
dehnert
|
c5da67d6cf
|
refined warning for automatic switch to policy iteration in exact mode
|
8 years ago |
dehnert
|
8cdbf281fa
|
make minmax solvers use policy iteration when --exact is set and no other method was explicitly set
|
8 years ago |
TimQu
|
9ca14a54fc
|
templated the LpSolvers
|
8 years ago |
TimQu
|
f46e8bcccf
|
fixed selecting LPMinMaxSolver in --exact mode
|
8 years ago |
TimQu
|
89f1796c56
|
Fixed creation of LpMinMaxSolver with the generalMinMaxSolverFactory
|
8 years ago |
TimQu
|
39549f6ebd
|
Moved some functionality of StandardMinMaxSolver into a subclass
|
8 years ago |
TimQu
|
49713eea72
|
Added new MinMaxMethod: 'acyclic' which potentially increases performance on acyclic mdps
|
8 years ago |
TimQu
|
2f49255db6
|
Improved storage::Scheduler. We can now consider arbitrary finite memory schedulers, potentially employing randomization.
|
8 years ago |
TimQu
|
536b1669c3
|
fixes for dtmc parameter lifting
|
8 years ago |
TimQu
|
ac6694f103
|
Improved sparse mdp model checking: Now allows hints for expected rewards
|
8 years ago |
dehnert
|
c467fa5f38
|
printing -1 as infinity for rational numbers and added clipping result to valid range where appropriate
|
8 years ago |
Sebastian Junges
|
d246517757
|
removed src prefix in all includes
|
9 years ago |
Sebastian Junges
|
e1d201c85e
|
c++ code compiles again after rename
|
9 years ago |
Sebastian Junges
|
3a7ee7867b
|
rename files (does not compile)
|
9 years ago |
TimQu
|
53ff482947
|
Caching of solvers can now be enabled/disabled
Former-commit-id: 498622f45c [formerly 1713f554a0 ]
Former-commit-id: 1015c7fef8
|
9 years ago |
TimQu
|
965d5b5e75
|
changed handling of auxiliary memory in MinMaxSolver
Former-commit-id: 602833f39d
|
9 years ago |
dehnert
|
2a90b5d737
|
more transient assignments
Former-commit-id: a18c4030eb [formerly 2f39009d01 ]
Former-commit-id: e38777fd53
|
9 years ago |
sjunges
|
548ba8bbeb
|
somehow managed my way through the policy guessing, several minor extensions to solvers
Former-commit-id: c4bb6453e7
|
9 years ago |