LIPIcs.STACS.2020.9.pdf
- Filesize: 0.51 MB
- 18 pages
The article investigates the relation between three well-known hypotheses. - H_{union}: the union of disjoint ≤^p_m-complete sets for NP is ≤^p_m-complete - H_{opps}: there exist optimal propositional proof systems - H_{cpair}: there exist ≤^{pp}_m-complete disjoint NP-pairs The following results are obtained: - The hypotheses are pairwise independent under relativizable proofs, except for the known implication H_{opps} ⇒ H_{cpair}. - An answer to Pudlák’s question for an oracle relative to which ¬H_{cpair}, ¬H_{opps}, and UP has ≤^p_m-complete sets. - The converse of Köbler, Messner, and Torán’s implication NEE ∩ TALLY ⊆ coNEE ⇒ H_{opps} fails relative to an oracle, where NEE =^{df} NTIME(2^O(2ⁿ)). - New characterizations of H_{union} and two variants in terms of coNP-completeness and p-producibility of the set of hard formulas of propositional proof systems.
Feedback for Dagstuhl Publishing