Why, intuitively, are propositional resolution proofs so long?

100 Views Asked by At

I'm trying to gain an intuitive understanding of why propositional resolution proofs tend to be so long.

As every essential prime implicant can be produced via resolution, intuitively I would have guessed that lower bounds on the length of resolution would be left open, or perhaps even equivalent to whether NP = co-NP. However, it is proven that resolution proofs can have exponential lower bounds, as shown by Tseitin's and Haken's work on the pigeonhole principle.

This establishes resolution as being a lot less useful than one would hope, and indeed this is so well known that modern solvers look at all manner of other refutation proofs.

Trying a resolution proof on a pigeonhole principle formula quickly yields an explosion of mostly useless clauses with tons of literals, but I don't exactly understand why this is the case.

Is there some ultra-simple example that can illustrate where resolution goes wrong, or provide some intuition for why auxiliary variables can (at least sometimes) rectify this issue?