• DocumentCode
    2144187
  • Title

    Hard sets are hard to find

  • Author

    Buhrman, H. ; van Melkebeek, D.

  • Author_Institution
    CWI, Amsterdam, Netherlands
  • fYear
    1998
  • fDate
    15-18 Jun 1998
  • Firstpage
    170
  • Lastpage
    181
  • Abstract
    We investigate the frequency of complete sets for various complexity classes within EXP under several polynomial-time reductions in the sense of resource bounded measure. We show that these sets are scarce: The sets that are complete under ⩽(nα-tt -)P reductions for NP, the levels of the polynomial-time hierarchy, and PSPACE have p2-measure zero for any constant α<1. The ⩽(nc-T)P-complete sets for EXP have p2 -measure zero for any constant c. Assuming MA≠EXP, the ⩽ ttP-complete sets for EXP have p-measure zero. A key ingredient is the Small Span Theorem, which states that for any set A in EXP at least one of its lower span (i.e., the sets that reduce to A) or its upper span (i.e., the sets that A reduces to) has p2-measure zero. Previous to our work, the theorem was only known to hold for ⩽bttp-reductions. We establish it for ⩽(n0(1)-tt)p-reductions
  • Keywords
    computational complexity; NP; PSPACE; Small Span Theorem; complete sets; complexity classes; polynomial-time hierarchy; polynomial-time reductions; resource bounded measure; Complexity theory; Computer science; Frequency; Polynomials; Uniform resource locators;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computational Complexity, 1998. Proceedings. Thirteenth Annual IEEE Conference on
  • Conference_Location
    Buffalo, NY
  • ISSN
    1093-0159
  • Print_ISBN
    0-8186-8395-3
  • Type

    conf

  • DOI
    10.1109/CCC.1998.694602
  • Filename
    694602