PRISM

Benchmark
Model:consensus v.1 (MDP)
Parameter(s)N = 4, K = 4
Property:steps_min (exp-reward)
Invocation (default)
./fix-syntax ./prism -javamaxmem 11g -cuddmaxmem 4g -ii -e 5e-2 -heuristic speed -ddextraactionvars 100 -maxiters 1000000 consensus.4.prism consensus.props --property steps_min -const K=4
Execution
Walltime:25.643999338150024s
Return code:0
Relative Error:0.024749952189064194
Log
PRISM
=====

Version: 4.5.dev
Date: Sat Mar 14 15:28:23 UTC 2020
Hostname: e72bdd194fc5
Memory limits: cudd=4g, java(heap)=11g
Command line: prism -javamaxmem 11g -cuddmaxmem 4g -ii -e 5e-2 -heuristic speed -ddextraactionvars 100 -maxiters 1000000 consensus.4.prism consensus.props --property steps_min -const K=4

Parsing model file "consensus.4.prism"...

Type:        MDP
Modules:     process1 process2 process3 process4 
Variables:   counter pc1 coin1 pc2 coin2 pc3 coin3 pc4 coin4 

Parsing properties file "consensus.props"...

5 properties:
(1) "c1": P>=1 [ F "finished" ]
(2) "c2": Pmin=? [ F "finished"&"all_coins_equal_1" ]
(3) "disagree": Pmax=? [ F "finished"&!"agree" ]
(4) "steps_max": R{"steps"}max=? [ F "finished" ]
(5) "steps_min": R{"steps"}min=? [ F "finished" ]

---------------------------------------------------------------------

Model checking: "steps_min": R{"steps"}min=? [ F "finished" ]
Model constants: K=4

Warning: Switching to sparse engine and (backwards) Gauss Seidel (default for heuristic=speed).

Warning: Switching to explicit engine to allow interval iteration on Rmin operator.

Building model...
Model constants: K=4

Computing reachable states... 43136 states
Reachable states exploration and model construction done in 0.712 secs.
Sorting reachable states list...

Time for model construction: 1.005 seconds.

Type:        MDP
States:      43136 (1 initial)
Transitions: 144352
Choices:     115840
Max/avg:     4/2.69
Building reward structure...

Starting expected reachability (min)...
Starting Prob1 (max)...
Prob1 (max) took 57 iterations and 0.231 seconds.
target=64, inf=0, rest=43072
For Rmin, checking for zero-reward ECs...
Time for checking for zero-reward ECs: 0.057 seconds, no zero-reward ECs found, proceeding normally.
Computing upper bound(s) for Rmin using the Dijkstra Sweep for Monotone Pessimistic Initialization algorithm...
Calculating incoming choices relation for Markov decision process...  done (0.053 seconds)
Time for computing upper bound(s) for Rmin using the DSI-MP algorithm: 0.431 seconds.
Upper bound for min expectation (Dijkstra Sweep MPI): 294914.0
Starting Prob1 (min)...
Prob1 (min) took 66 iterations and 0.202 seconds.
Relevant sub-MDP is contracting, proceed...
Starting interval iteration (min, with Power method)...
Iteration 1238: max relative diff=78.0896956402, 5.01 sec so far
Iteration 2492: max relative diff=9.09036474794, 10.01 sec so far
Iteration 3753: max relative diff=1.17605860758, 15.01 sec so far
Iteration 5016: max relative diff=0.153943829348, 20.01 sec so far
Max relative diff between upper and lower bound on convergence: 0.0499371794777
Interval iteration (min, with Power method) took 5715 iterations, 1649211840 multiplications and 22.813 seconds.
Maximum finite value in solution vector at end of interval iteration: 795.1922203802607
Expected reachability took 23.787 seconds.

Value in the initial state: 787.0079632812013

Time for model checking: 23.854 seconds.

Result: 787.0079632812013 (value in the initial state)


Overall running time: 25.473 seconds.

---------------------------------------------------------------------

Note: There were 2 warnings during computation.