$BENCH_HOME/bin/storm-pomdp --prism $BENCH_HOME/models/clean/clean.prism --prop $BENCH_HOME/models/clean/clean.props rbrmax2 -const N=6,B1=60,B2=5 --timemem --statistics --revised --reward-aware 4,0 --belief-exploration unfold --size-threshold 524288
Storm-pomdp. Sequential approach, cost aware, with cutoffs and size threshold 2^19
Storm-POMDP 1.9.1 (dev)
Date: Mon Feb 10 10:23:17 2025
Command line arguments: --prism $BENCH_HOME/models/clean/clean.prism --prop $BENCH_HOME/models/clean/clean.props rbrmax2 -const 'N=6,B1=60,B2=5' --timemem --statistics --revised --reward-aware '4,0' --belief-exploration unfold --size-threshold 524288
Current working directory: $BENCH_HOME/experiments64gb
Time for model input parsing: 0.001s.
Time for model construction: 0.008s.
--------------------------------------------------------------
Model type: POMDP (sparse)
States: 37
Transitions: 74
Choices: 50
Observations: 2
Reward Models: energy, clean
State Labels: 3 labels
* deadlock -> 0 item(s)
* init -> 1 item(s)
* goal -> 1 item(s)
Choice Labels: 3 labels
* clean -> 13 item(s)
* move -> 13 item(s)
* consume -> 24 item(s)
--------------------------------------------------------------
Analyzing property 'Pmax=? [true U^{rew{"energy"}<=60 , rew{"clean"}>5 }"goal"]'
Perform unfolding for observation levels.
bounded reachability processing done. POMDP Information:
--------------------------------------------------------------
Model type: POMDP (sparse)
States: 338
Transitions: 787
Choices: 508
Observations: 5
Reward Models: dim0_levelReward
State Labels: 4 labels
* goal -> 35 item(s)
* dim1_active -> 5 item(s)
* deadlock -> 0 item(s)
* init -> 1 item(s)
Choice Labels: 3 labels
* move -> 170 item(s)
* consume -> 168 item(s)
* clean -> 170 item(s)
--------------------------------------------------------------
Transformed formula: Pmax=? [true Urew{"dim0_levelReward"}<=15 ("goal" & "dim1_active")]
Time for pre-processing: 0.000s.
Exploring the belief MDP...
Exploring the belief space...
Exploration stopped before all beliefs were explored. 524289 beliefs discovered. 475383 beliefs explored.
Constructing the belief MDP...
--------------------------------------------------------------
Model type: MDP (sparse)
States: 524290
Transitions: 1133040
Choices: 789492
Reward Models: dim0_levelReward
State Labels: 3 labels
* target -> 23 item(s)
* init -> 1 item(s)
* bottom -> 1 item(s)
Choice Labels: none
--------------------------------------------------------------
Analyzing property 'Pmax=? [true Urew{"dim0_levelReward"}<=15 "target"]' on the belief MDP...
Transformation of transition rewards resulted in a model with 717996 states. 1.369463465 times more states than the original belief MDP.
Merging of sink states resulted in a model with 188125 states.
---------------------------------
Statistics:
---------------------------------
#checked epochs: 17.
overall Time: 0.278s.
Epoch Model building Time: 0.190s.
Epoch Model checking Time: 0.087s.
---------------------------------
Time for exploring beliefs: 2.623s.
Time for building the belief MDP: 0.259s.
Time for analyzing the belief MDP: 0.761s.
##### POMDP Approximation Statistics ######
# Input model:
--------------------------------------------------------------
Model type: POMDP (sparse)
States: 338
Transitions: 787
Choices: 508
Observations: 7
Reward Models: dim0_levelReward
State Labels: 4 labels
* goal -> 35 item(s)
* dim1_active -> 5 item(s)
* deadlock -> 0 item(s)
* init -> 1 item(s)
Choice Labels: 3 labels
* move -> 170 item(s)
* consume -> 168 item(s)
* clean -> 170 item(s)
--------------------------------------------------------------
# Max. Number of states with same observation: 168
# Total check time: 3.948s
##########################################
Result: ≥ 0.9260842476
Time for POMDP analysis: 4.006s.
Performance statistics:
* peak memory usage: 809MB
* CPU time: 3.814s
* wallclock time: 4.017s
############################## Notes ##############################
Storm-pomdp. Sequential approach, cost aware, with cutoffs and size threshold 2^19