Solovay functions and K-triviality

Authors Laurent Bienvenu, Wolfgang Merkle, Andre Nies

Thumbnail PDF


  • Filesize: 0.6 MB
  • 12 pages

Document Identifiers

Author Details

Laurent Bienvenu
Wolfgang Merkle
Andre Nies

Cite AsGet BibTex

Laurent Bienvenu, Wolfgang Merkle, and Andre Nies. Solovay functions and K-triviality. In 28th International Symposium on Theoretical Aspects of Computer Science (STACS 2011). Leibniz International Proceedings in Informatics (LIPIcs), Volume 9, pp. 452-463, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2011)


As part of his groundbreaking work on algorithmic randomness, Solovay demonstrated in the 1970s the remarkable fact that there are computable upper bounds of prefix-free Kolmogorov complexity $K$ that are tight on infinitely many values (up to an additive constant). Such computable upper bounds are called Solovay functions. Recent work of Bienvenu and Downey~[STACS 2009, LIPIcs 3, pp 147-158] indicates that Solovay functions are deeply connected with central concepts of algorithmic randomness such as $Omega$ numbers, K-triviality, and Martin-Loef randomness. In what follows, among other results we answer two open problems posed by Bienvenu and Downey about the definition of $K$-triviality and about the Gacs-Miller-Yu characterization of Martin-Loef randomness. The former defines a sequence A to be K-trivial if K(A|n) <=^+ K(n), the latter asserts that a sequence A is Martin-Loef random iff C(A|n) >=^+ n-K(n). So both involve the noncomputable function K. As our main results we show that in both cases K(n) can be equivalently replaced by any Solovay function, and, what is more, that among all computable functions such a replacement is possible exactly for the Solovay functions. Moreover, similar statements hold for the larger class of all right-c.e. in place of the computable functions. These full characterizations, besides having significant theoretical interest on their own, will be useful as tools when working with K-trivial and Martin-Loef random sequences.
  • Algorithmic randomness
  • Kolmogorov complexity
  • K-triviality


  • Access Statistics
  • Total Accesses (updated on a weekly basis)
    PDF Downloads