Replaced isLess in DFTModelChecker

[volkm]

Addresses some parts of #831

[sjunges]

LGTM, but why not use isLess with precision (or rather: tolerance) 0?

[volkm]

My main reason was that isLess requires a ConstantsComparator. But I agree that it would be more clear to use isLess or even better isLessEqual.
Maybe a more general question: when should we use constants.h and when ConstantsComparator? It seems that both overlap in some functionality.

[sjunges]

I would say that the comparator is only used when you have a state, i.e., the type of precision.

Maybe we should extend the constants to have a definition of isLess that does not require any precision?

[volkm]

I agree with your idea to add isLess (without precision) to constants.h. I would simply use the < operator here.
I took a look at ElementLess which we already define and which also uses std::less for comparison. I suggest to remove the double specialization though because it has two issues: a magic number for precision and it only works for non-negative numbers.

[sjunges]

I think we haven't used < because the code has to also work for polynomials etc and the intended semantics of the < operator is not clear?