dehnert
|
99d9a9710d
|
Further steps to make everything work again.
Former-commit-id: 3f45a49dab
|
11 years ago |
dehnert
|
7ec3e8b214
|
Further fixes for new variable handling. libstorm now compiles again, yay.
Former-commit-id: a9ac5c0356
|
11 years ago |
dehnert
|
f76d0f93eb
|
Adapted LP solver interface to new variable handling.
Former-commit-id: c2a7eea5db
|
11 years ago |
dehnert
|
7ea6ec3644
|
Further refactoring.
Former-commit-id: 7769132a68
|
11 years ago |
dehnert
|
bdfbc50dab
|
Removed some superfluous stuff.
Former-commit-id: ca742ac658
|
11 years ago |
dehnert
|
92d550be12
|
More and more refactoring.
Former-commit-id: b2f5b25c92
|
11 years ago |
dehnert
|
398f6c4e86
|
Partly adapted code to new 'type system'.
Former-commit-id: 6c200df5f0
|
11 years ago |
dehnert
|
983a7d78c2
|
Further work on expressions.
Former-commit-id: 4774f0136d
|
11 years ago |
dehnert
|
fff18f2789
|
Intermediate commit (refactoring expressions).
Former-commit-id: e7cc7e95c2
|
11 years ago |
dehnert
|
809217c359
|
Refactored some parts of expressions. In particular, visitors now can return anything they want by using boost::any.
Former-commit-id: 0f6af138ae
|
11 years ago |
dehnert
|
b5d55335a6
|
All tests passing again.
Former-commit-id: ffa8bef2d2
|
11 years ago |
dehnert
|
90b0f20167
|
Reachability Rewards can now be computed in parametric DTMCs (modulo bugs)
Former-commit-id: 26ee20ef76
|
11 years ago |
dehnert
|
554287e082
|
Fixed minor issue that caused problems with the measure-driven initial partition and rewards.
Former-commit-id: 7379da548d
|
11 years ago |
dehnert
|
b7492d543a
|
Further work regarding rewards in parameterized models. Note: this includes some debug output.
Former-commit-id: ac65f020a5
|
11 years ago |
dehnert
|
7d0ae06f9f
|
Fixed creation of empty blocks under certain circumstances in bisimulation.
Former-commit-id: f1240e234b
|
11 years ago |
dehnert
|
cca4ba4ecf
|
Removed debug time measurements.
Former-commit-id: 17cdf5c41c
|
11 years ago |
dehnert
|
0bc685969d
|
Moved from call to list::size to counting member in bisimulation partition to avoid gcc's O(n) list::size.
Former-commit-id: aaae9886b7
|
11 years ago |
dehnert
|
0ad4c5f867
|
More debug times.
Former-commit-id: fb6d8c06f7
|
11 years ago |
dehnert
|
9b91d388b7
|
Even morer debug times.
Former-commit-id: 843fcf4313
|
11 years ago |
dehnert
|
f476caf62e
|
More debug timings.
Former-commit-id: 07ffbf3fcd
|
11 years ago |
dehnert
|
0af2b8d148
|
More debug stats.
Former-commit-id: 1885b7ff67
|
11 years ago |
dehnert
|
8c403628f2
|
Added some debug statistics to bisim.
Former-commit-id: 6a93014021
|
11 years ago |
dehnert
|
4b8f2e7a0b
|
Next splitter is now chosen more deterministically.
Former-commit-id: 7f92208d1c
|
11 years ago |
dehnert
|
894c3bb497
|
Added missing header.
Former-commit-id: 30d5444d23
|
11 years ago |
dehnert
|
b305a3b498
|
Switched to FactorizedPolynomial as the basis for rational functions and added missing reward construct for one NAND model.
Former-commit-id: 8bb62ee1d2
|
11 years ago |
dehnert
|
7014d289e8
|
Fixed some issues related to bisimulation in the presence of state rewards.
Former-commit-id: 7f26a7bcf9
|
11 years ago |
dehnert
|
61e78f8d12
|
Adapted parameterized NAND example to use state rewards instead of transition rewards. Also, the unfactorized polynomials are now used to build and compute everything. We should detect cyclic models and use the factorized polynomials for them.
Former-commit-id: c4179f2029
|
11 years ago |
dehnert
|
08959a6a32
|
Intermediate commit.
Former-commit-id: d3c8fe1b9b
|
11 years ago |
dehnert
|
c4c7794069
|
Intermediate commit.
Former-commit-id: 19002ec2c1
|
11 years ago |
dehnert
|
a7bce9e520
|
Removed debug output and fixed the reward issue a bit more.
Former-commit-id: ecbbeff14e
|
11 years ago |
dehnert
|
7cd0dfe8b0
|
Fixed an issue regarding the reward model generation.
Former-commit-id: 237acf99f9
|
11 years ago |
dehnert
|
1b4d2a92db
|
Started working on making bisimulation work for models with (state-based) rewards.
Former-commit-id: b1029210f6
|
11 years ago |
dehnert
|
370a0ae476
|
Fixed some issues in bisimulation and added some tests.
Former-commit-id: 98801de9db
|
11 years ago |
dehnert
|
2f20abf47f
|
The user can now select on the command line which reward model of a symbolic model is to be used (as a second [optional] argument to --symbolic).
Former-commit-id: 02f998e5dd
|
11 years ago |
PBerger
|
9fc68a554c
|
Cherry-picked a fix for GCC from branch.
Former-commit-id: 98f7c52b34
|
11 years ago |
dehnert
|
23c6d14426
|
Replaced inline conversions to explicit conversions in an attempt to prevent gcc from using uninitialized values when using chrono.
Former-commit-id: de7dc5dbc6
|
11 years ago |
dehnert
|
d3721196c4
|
Typed return value of lambda instead of using 'auto'.
Former-commit-id: 03c3a38b9c
|
11 years ago |
dehnert
|
f3048d31c2
|
Small bugfix for bisimulation decomposition.
Former-commit-id: eae1447df4
|
11 years ago |
dehnert
|
e6904dcb21
|
Renamed bisimulation decomposition class to reflect that now also weak bisimulations can be computed.
Former-commit-id: 1a654b7110
|
11 years ago |
dehnert
|
f90ac5c8c3
|
First working version of weak bisimulation for DTMCs.
Former-commit-id: 8a7d76de4f
|
11 years ago |
dehnert
|
7257bb23c3
|
Further work on weak bisimulation. Model checking can now be done from tne command line again.
Former-commit-id: 5f338260e6
|
11 years ago |
dehnert
|
391f3225e4
|
Added unparameterized NAND example. Further work on weak bisimulation.
Former-commit-id: 0936743f1e
|
11 years ago |
dehnert
|
5bc593174e
|
Further work on weak bisimulation.
Former-commit-id: 3ad48ee0a3
|
11 years ago |
dehnert
|
56aec18a48
|
Added bisimulation settings. Further work on weak bisimulation.
Former-commit-id: c04759575a
|
11 years ago |
dehnert
|
97158ee72e
|
Started on weak bisimulation.
Former-commit-id: 595caab54e
|
11 years ago |
dehnert
|
754e168ace
|
Bugfix for bisimulation.
Former-commit-id: da93a5d4db
|
11 years ago |
dehnert
|
d3fc2d8fbf
|
Fixed small but important bug in SCC decomposition that led to wrong results when using MSVC.
Former-commit-id: 07358dc2e8
|
11 years ago |
dehnert
|
08ac566db2
|
Corrected typedef. Clang and gcc should now also be fine under Linux.
Former-commit-id: 46f8d43d47
|
11 years ago |
dehnert
|
74351f9884
|
Switched from const_iterator to iterator in bisimulation to make stdlibc++ happy (libc++ is already happy, though).
Former-commit-id: 37fc55d0cf
|
11 years ago |
dehnert
|
3dfc6a7b74
|
Pimped bisimulation a bit.
Former-commit-id: a27ea8b996
|
11 years ago |