Functions Computable with Nonadaptive Queries to NP

  • Authors:

  • H. Buhrman;J. Kadin;T. Thierauf

  • Affiliations:

  • CWI, PO Box 94079, 1090 GB Amsterdam, The Netherlands buhrman@cwi.nl, The Netherlands;Department of Computer Science, University of Maine, 5752 Neville Hall, Orono, ME 04469-5752, USA, USA;Abteilung Theoretische Informatik, Universität Ulm, Oberer Eselsberg, 89069 Ulm, Germany thierauf@informatik.uni-ulm.de, Germany

  • Venue:
  • Theory of Computing Systems
  • Year:
  • 1998

Quantified Score

Hi-index 0.00

Visualization

Abstract

We study FP||NP , the class of functions that can be computed in polynomial time with nonadaptive queries to an NP oracle. This is motivated by the question of whether it is possible to compute witnesses for NP sets within FP||NP . The known algorithms for this task all require sequential access to the oracle. On the other hand, there is no evidence known yet that this should not be possible with parallel queries.We define a class of optimization problems based on NP sets, where the optimum is taken over a polynomially bounded range (NPbOpt). We show that if such an optimization problem is based on one of the known NP-complete sets, then it is hard for FP||NP . Moreover, we characterize FP||NP as the class of functions that reduces to such optimization functions. We call this property strong hardness.The main question is whether these function classes are complete for FP||NP . That is, whether it is possible to compute an optimal value for a given optimization problem in FP||NP . We show that these optimization problems are complete for FP||NP , if and only if one can compute membership proofs for NP sets in FP||NP . This indicates that the completeness question is a hard one.