Assertion with zero and propagation engine about boolector

[mpreiner]

I'm not sure I understand what the issue is? Can you please elaborate.

[gsingla21]

with PROP Engine ON, Incremental Mode On,
(x>=1 && x<=10) provides different possible value of x (e.g. 1,9,8,6,...) everytime boolector_sat is executed.
But for
(x>=0 && x<=10) provides only x=0 for every time boolector_sat is executed. (e.g. 0,0,0,0,0,...)
Not just this one, problem lies with every constraint where x=0 is a possible solution. It doesn't propagate if x=0 is a possible solution.
As you can see in above table, x>=0 equivalent btor file is just zero, it should have been something like
1 var 32 x
2 zero 32
3 ult 1 1 2
4 root 1 -3

[mpreiner]

Just to be clear: the prop engine is a local search engine, which can be randomized via the BTOR_OPT_SEED option. It is not guaranteed that it will always produce different model values for subsequent boolector_sat calls. If you want Boolector to provide different model values, you have to explicitly block these model values with new constraints. It is just coincidence that in your case it will return different values.

Your example code for (x >= 1 && x <= 10) produces the following for me:

i: 0 sat x: 15
i: 1 sat x: 15
i: 2 sat x: 15
i: 3 sat x: 15
i: 4 sat x: 15
i: 5 sat x: 5
i: 6 sat x: 15
i: 7 sat x: 15
i: 8 sat x: 15
i: 9 sat x: 15

Using a different random seed it will produce:

i: 0 sat x: 15
i: 1 sat x: 7
i: 2 sat x: 15
i: 3 sat x: 15
i: 4 sat x: 15
i: 5 sat x: 5
i: 6 sat x: 15
i: 7 sat x: 15
i: 8 sat x: 8
i: 9 sat x: 15

For the x >= 0 case it always returns 0 since this is the first value the local search engine tries and succeeds.

As you can see in above table, x>=0 equivalent btor file is just zero, it should have been something like
1 var 32 x
2 zero 32
3 ult 1 1 2
4 root 1 -3

When Boolector prints

1 zero 1
2 root 1 -1

just means that Boolector's preprocessor+rewriting was able to simplify the formula to true. This has nothing to do with the propagation engine.

[gsingla21]

Got it, Thanks

[valeryjust]

@mpreiner @gsingla21 Sorry for bothering you. I want to know is there a way to change the first value the local search engine tries and succeeds ?