TimQu
fcd277c42a
added an option that enables building of state valuations. Also include the state valuations when the model is exported to .dot format
7 years ago
dehnert
3c844a487f
some more optimizations
7 years ago
sjunges
10b1f65840
eliminating selfloops in pomdps
8 years ago
sjunges
b120b74fa9
StateActionPair to index should be part of nondeterministicmodel
8 years ago
sjunges
6e506e5a66
moved application of permissive scheduler to an own transformer
8 years ago
Sebastian Junges
fb8dd88314
fixes in parsing, support for POMDPs in DRN
8 years ago
dehnert
ad456916e9
first working version of sparse reward model quotienting
8 years ago
dehnert
722cb3109c
dd quotient extraction of reward models in dd bisimulation
8 years ago
dehnert
34e23f94fc
started on reward model preservation in DD bisimulation
8 years ago
dehnert
f96403de9e
added reduction to state-based rewards to symbolic (reward) models
8 years ago
Sebastian Junges
09dcb149c2
storm parses and builds pomdps
8 years ago
Sebastian Junges
63fd02937f
pomdps
8 years ago
sjunges
bbb4d16965
towards support for pomdps in storm
8 years ago
sjunges
bc3646604d
Extend ModelType with an entry for POMDPs
8 years ago
dehnert
d23547d99f
started optimizing some DdManager methods
8 years ago
dehnert
2b0911d627
more work on MDP bisimulation
8 years ago
dehnert
a1db269e8f
started on debugging MDP bisimulation
8 years ago
dehnert
f3ebfaa90f
more work on MDP bisimulation
8 years ago
dehnert
b25ef3f09c
introduced symbolic bisimulation modes lazy and eager, fixed bug in sparse quotient extraction
8 years ago
TimQu
e8e189723f
fixed applying memoryless schedulers
8 years ago
TimQu
5651b23771
fixing minor compiling issue
8 years ago
TimQu
a348f6ea8e
function to apply a given scheduler to a nondeterministic model
8 years ago
TimQu
19925ac74d
implemented value iteration based Long run average rewards for Markov automata by Butkova et al. (TACAS 2017)
8 years ago
dehnert
ea02ea0838
started overhaul of cli/api
8 years ago
TimQu
e7a8357ee6
Fixed some tests
8 years ago
TimQu
576f92568e
StateValuations and ChoiceOrigins are now members of a sparse::Model.
A model can now be constructed by providing a modelComponents struct.
8 years ago
dehnert
f0f4cd7390
first version of sparse quotient extraction for dd bisimulation
8 years ago
TimQu
464bdc389c
improved state valuations class
8 years ago
TimQu
722e67fe64
parsing choice labels for explicit models
8 years ago
TimQu
1d329176ba
Resolved compiling issues due to recent merge
8 years ago
TimQu
cd5ee63cce
fixed preserving the choice labeling when an ma is closed
8 years ago
TimQu
b531dccad9
.dot output for deterministic models with choice labels
8 years ago
TimQu
e7bc5fdef9
fixed several minor bugs regarding the choicelabeling
8 years ago
TimQu
bf97d79573
moved building the choice origin strings into the ChoiceOrigins class
8 years ago
TimQu
db31c1cb11
improved .dot export of models with choice labeling
8 years ago
TimQu
6537fd8b72
Replaced the old choice labeling with the new one and used choice origins for the minimal command set counterexample generators
8 years ago
TimQu
759e351e95
Improved explicit model building:
- There is now an option to generate a choice labeling that corresponds to the specified action names.
- The old choice labeling (where each choice was labeled with an index set representing the corresponding prism commands) is renamed to choiceOrigins and has been improved towards support of other input formats (such as Jani) and other applications such as scheduler synthesis
8 years ago
TimQu
4413afb542
used new helper functions at some points in the code
8 years ago
dehnert
7f346d2f0b
more work on quotient extraction
8 years ago
Sebastian Junges
5c7d3db743
towards proper side constraints for parametetric systems
8 years ago
TimQu
748e100aad
fixed/improved .dot output for MAs and Mdps. We now also display the index of each choice.
8 years ago
TimQu
6598ade4ac
fix for getting the choices with zero reward
8 years ago
TimQu
c14213f9a6
Reward model can now retrieve the set of choices with zero reward
8 years ago
Sebastian Junges
a2ed0fc4bf
item labelling class
8 years ago
Sebastian Junges
586929ea64
As we do not support windows, we can also get rid of:
#ifndef WINDOWS
especially since the guards were around move-constructors, which are supported under Windows since Visual Studio 2015
8 years ago
dehnert
03ad4c2783
first version of symbolic bisimulation minimization
8 years ago
TimQu
09552d43a3
IsExact number trait
8 years ago
dehnert
853b035473
fixed bug and added testsfor symbolic linear equation solver (rational number and rational function)
8 years ago
TimQu
dd40254628
PLA for continuous models
8 years ago
dehnert
952776a057
hybrid engine working for rational numbers
8 years ago