tag:blogger.com,1999:blog-25562705.post3233634732949089193..comments2022-09-15T22:27:21.570-04:00Comments on Adventures in Computation: Grim TriggerAaronhttp://www.blogger.com/profile/09952936358739421126noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-25562705.post-48473595916387128162008-09-26T12:51:00.000-04:002008-09-26T12:51:00.000-04:00SAT is (not) NP-completehttp://arxiv.org/abs/0806....SAT is (not) NP-complete<BR/>http://arxiv.org/abs/0806.2947<BR/>Here is a easy-to-understand for this:<BR/>Let L be the Prolog program:<BR/>Fact:<BR/>Age-About-21(John,0.9).<BR/>Goal:<BR/>?- Age-About-21(John,0.9).<BR/>The Prolog system answers two different contradictory answers:<BR/>"Yes" = "1" and "0.9" in the fact.<BR/>Try the goal:<BR/>?- Age-About-21(John,0.3).<BR/>Again, the system answers two contradictory answers:"No" = "0" and "0.9".<BR/>Clearly, this language is in P, but how to reduce it to SAT?<BR/>How to reduce instances of the FLP problem whose output is two contradictory truth values to the SAT problem whose output is only one.<BR/>It is easy to show that ZFC is inconsistent via 2 (independent) proofs.Anonymousnoreply@blogger.com