Complexity for type-2 relations

26Citations
Citations of this article
7Readers
Mendeley users who have this article in their library.

Abstract

If is an alphabet, then a type-2 functional is a (partial) function whose arguments are elements of and functions from into. Type-2 relations are the domains of such functionals. We consider the natural extension, POLY, of the class of polynomial time functions to include type-2 functionals, and a variant, POLY, in which the time bounds depend only on the string arguments. Using these, we define two possible extensions of the (relativized) polynomial hierarchy to include type-2 relations. For example, Ef = Il£ is the class of relations whose characteristic functionals are in POLY. Then Ef+1 is the class of relations definable by POLY length bounded existential quantification of relations in IIf, and dually for nf+1. Thus Ef is the type-2 analogue of NP. For any function g, we may relativize the definitions of Ef (Ilf) to obtain Ef* (Ilf*). Let Ef denote Ef*.g Some properties of the (relativized) hierarchy are studied. A similar analysis is carried out for the hierarchy based on POLY. In addition, we consider some topological notions that seem ‘naturally’ associated with time and space bounded computations of oracle Turing machines, and we give topological characterizations of several classes of type-2 relations. In particular, we give a topological characterization of Ef. We use these characterizations to examine analogues of several well-known open questions of computational complexity theory. For example, we show that a certain type-2 analogue of the NP = PSP ACE question has a negative answer. These results suggest that topological considerations are an integral part of the study of resource bounded computations of oracle Turing machines. © 1990 by the University of Notre Dame. All rights reserved.

Cite

CITATION STYLE

APA

Townsend, M. (1990). Complexity for type-2 relations. Notre Dame Journal of Formal Logic, 31(2), 241–262. https://doi.org/10.1305/ndjfl/1093635419

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free