A-lemma-on-a-total-function-defined-over-the-Baker–Gill–Solovay-set-of-polynomial-Turing-machines.pdfVIP

A lemma on a total function defined over the Baker–Gill–Solovay set of polynomial Turing machines.∗ N. C. A. da Costa F. A. Doria Institute for Advanced Studies, University of S˜ao Paulo. Av. Prof. Luciano Gualberto, trav. J, 374. 05655–010 S˜ao Paulo SP Brazil. ncacosta@usp.br doria1000@.br doria@lncc.br May 25, 2006 Abstract We prove here a lemma that connects some properties of the so–called “counterexample function” to the P = NP conjecture over the Baker– Gill–Solovay set of polynomial Turing machines to the behavior of the same function “at large,” over the set of all polynomial Turing machines. ∗Partially supported by FAPESP and CNPq. Alternative address for F. A. Doria: Re- search Center for Mathematical Theories of Communication and Program IDEA, School of Communications, Federal University at Rio de Janeiro, Av. Pasteur, 250. 22295–900 Rio RJ Brazil. 1 1 Introduction We deal here with a property of the Baker–Gill–Solovay (BGS) set [1] of poly- nomial Turing machines. The BGS set is a kind of ‘representation set’ for poly- nomial machines in the following sense: for every computable function f with a polynomial algorithm (a polynomial Turing machine), there is one such polyno- mial algorithm for f in the BGS set, and there are only polynomial algorithms in BGS. Actually there are infinitely many polynomial algorithms for each f in BGS, but not every polynomial algorithm for f will be in BGS. The BGS set was conceived to rigorously formulate the P NP question (see below). Since the set of all polynomial Turing machines isn’t recursive in