TimQu
|
88ee0bbf67
|
RewardUnfolding: If statistics are enabled, Log when an acyclic epoch model is found.
|
7 years ago |
|
TimQu
|
5ad100e652
|
quantiles: Added some statistics.
|
8 years ago |
|
TimQu
|
fb7078770d
|
rewardbounded: Various fixes.
|
8 years ago |
|
TimQu
|
dd93b1dae9
|
rewardbounded: Improved code structure.
|
8 years ago |
|
TimQu
|
6aeb75e3bd
|
quantiles: Supporting two-dimensional quantiles with the same optimization direction of quantile bounds (max,max or min,min).
|
8 years ago |
|
TimQu
|
a99dd5e4d1
|
quantiles: Better code re-usage, better structure, support for 'open' and 'non-open' dimensions, single dimensional quantiles should work now.
|
8 years ago |
|
TimQu
|
1d5f2410b5
|
rewardBounded/RewardUnfolding: Allowed the case that not all dimensions have a bound a priori.
|
8 years ago |
|
TimQu
|
4ac23d630f
|
quantiles: Added support for formulas with trivial bounds (i.e., >=0).
|
8 years ago |
|
Tim Quatmann
|
6e8aef2acc
|
Checking formulas with >=0 bound.
|
8 years ago |
|
Tim Quatmann
|
f99a24acd2
|
more work on quantiles.
|
8 years ago |
|
Tim Quatmann
|
82402ba3ae
|
rewardbounded: Moved epoch model analysis to a separate file.
|
8 years ago |
|
TimQu
|
d796ee74de
|
(workplace switch)
|
8 years ago |
|
Tim Quatmann
|
d3abeb5f45
|
Started implementation on quantiles.
|
8 years ago |
|
TimQu
|
dc2654ce60
|
Quantiles: made the SparseMdpPrctlModelChecker call the QuantileHelper for quantile formulas
|
8 years ago |
|
TimQu
|
fceaeed557
|
Fixed a recently introduced issue with interval bounded ctmcs.
|
8 years ago |
Matthias Volk
|
230ac20480
|
Added progress measurements for Unif+ iterations and steps
|
8 years ago |
|
TimQu
|
629de20da0
|
Fixed running in an infinite loop when computing LRA on markov automata with relative precision.
|
8 years ago |
|
TimQu
|
0c905e2323
|
Fixed an issue where time-bounded properties were wrongly computed on a ctmc that only consists of goal-states.
|
8 years ago |
|
Matthias Volk
|
94a1d47103
|
Use IMCA method for bounded until with lower bound > 0
|
8 years ago |
|
Matthias Volk
|
01a23856a8
|
Drastically decreased memory consumption of Unif+
|
8 years ago |
|
Matthias Volk
|
d054f3c64a
|
Result for upper bounds needs only be calculated for k=0
|
8 years ago |
|
Matthias Volk
|
4c5b041340
|
Debug output
|
8 years ago |
|
Matthias Volk
|
6220b114b5
|
Small simplifications
|
8 years ago |
|
Matthias Volk
|
f9fb90499d
|
Only keep track of results from the last iteration (instead of all iterations) for 2 of the 3 vectors
|
8 years ago |
|
Matthias Volk
|
8ae800b130
|
Changed iteration order to iterate over stepsize in outer loop
|
8 years ago |
|
Matthias Volk
|
cbd6139613
|
Small changes
|
8 years ago |
|
Matthias Volk
|
b8b2c58dab
|
Started on some refactoring in Unif+
|
8 years ago |
|
Matthias Volk
|
6066ecd590
|
Added struct for Unif+ vectors
|
8 years ago |
|
Matthias Volk
|
70d3f8d811
|
Unified order of function arguments for Unif+
|
8 years ago |
|
TimQu
|
27d87da93d
|
Fixed value iteration based LRA method for Markov Automata, where end components do not contain probabilistic states.
|
8 years ago |
|
TimQu
|
c614e9d747
|
Fixed Value Iteration based LRA computation
|
8 years ago |
|
TimQu
|
985319c7dd
|
Tweaked LRA computation for MDPs and MAs in sound mode to meet precision requirements.
|
8 years ago |
Sebastian Junges
|
43688d09ea
|
reward infinity scheduler extraction is now correct
|
8 years ago |
|
Sebastian Junges
|
93ca559c83
|
additional sanity checks for scheduler extraction
|
8 years ago |
|
Sebastian Junges
|
16d7dccb4e
|
I am utterly stupid. Fixed an assertion that I changed yesterday
|
8 years ago |
|
Sebastian Junges
|
5d0ec15ad4
|
clarified error message, as the reward models are present (according to output) but simply empty
|
8 years ago |
|
Sebastian Junges
|
9a0794fca1
|
refined error message wrt unexpected type of scheduler
|
8 years ago |
|
TimQu
|
9be488b969
|
Enabling expected time queries for ctmcs in the hybrid engine.
|
8 years ago |
|
TimQu
|
e6fc962e5e
|
In exact mode, use LP as LRA Method for nondeterministic models.
|
8 years ago |
|
TimQu
|
e94b37d2f5
|
instantaneous reward properties for continuous time models can not be handled in exact mode.
|
8 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.
|
8 years ago |
|
TimQu
|
bb7f4dd6d7
|
canHandle method of exact continuous time model checkers now return 'false' for time-bounded properties
|
8 years ago |
|
TimQu
|
c388d1c8fe
|
making sure that functions in jani models and formulas in prism programs are substituted before flattening the model
|
8 years ago |
|
TimQu
|
b5bb87404c
|
replaced occurrences of 'substituteConstants' by 'substituteConstantsFormulas'
|
8 years ago |
dehnert
|
51be532695
|
pulled out parsing from abstraction-refinement classes
|
8 years ago |
|
TimQu
|
0332935451
|
Supporting TimeOperatorFormulas for MDPs and DTMCs in Sparse, Hybrid, and Dd engine
|
8 years ago |
Joachim Klein
|
d0e2d099bf
|
AbstractModelChecker: In error messages, include class name of the actual model checker
|
8 years ago |
|
Joachim Klein
|
da4dfc35a8
|
AbstractModelChecker: add getClassName()
|
8 years ago |
|
dehnert
|
abe7510ae9
|
added clearing requirements
|
8 years ago |
|
dehnert
|
9ed6f084e7
|
adding uniqueness constraint in LRA computation also for fixed-point formulation
|
8 years ago |
