• DocumentCode
    2891653
  • Title

    Simulating BPP using a general weak random source

  • Author

    Zuckerman, David

  • Author_Institution
    Lab. Comput. Sci., MIT, Cambridge, MA, USA
  • fYear
    1991
  • fDate
    1-4 Oct 1991
  • Firstpage
    79
  • Lastpage
    89
  • Abstract
    It is shown how to simulate BPP and approximation algorithms in polynomial time using the output from a δ-source. A δ-source is a weak random source that is asked only once for R bits, and must output an R-bit string according to some distribution that places probability no more than 2-δR on any particular string. Also given are two applications: one to show the difficulty of approximating the size of the maximum clique, and the other to the problem of implicit O(1) probe search
  • Keywords
    computational complexity; file organisation; polynomials; search problems; BPP simulation; R-bit string; approximation algorithms; general weak random source; maximum clique; polynomial time; probe search; weak random source; Application software; Approximation algorithms; Clocks; Computer science; Diodes; Entropy; History; Physics computing; Polynomials; Probes;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Foundations of Computer Science, 1991. Proceedings., 32nd Annual Symposium on
  • Conference_Location
    San Juan
  • Print_ISBN
    0-8186-2445-0
  • Type

    conf

  • DOI
    10.1109/SFCS.1991.185351
  • Filename
    185351