TOPICS
Search

Chaitin's Constant


Chaitin's constant, also called a Chaitin omega number, introduced by Chaitin (1975), is the halting probability of a prefix-free (self-delimiting) universal Turing machine that is optimal: simulation of any other prefix-free Turing machine increases program length by at most a constant independent of the program. Every Chaitin constant is simultaneously computably enumerable (the limit of a computable, increasing, converging sequence of rationals), and algorithmically random (its binary expansion is an algorithmically random sequence), hence noncomputable (Chaitin 1975).

A Chaitin constant can therefore be defined as

 Omega_U=sum_(p halts)2^(-|p|),
(1)

which gives the probability that a stream of independent fair random bits begins with a halting program for U, where |p| is the length in bits of program p.

The value of a Chaitin constant is highly machine-dependent. Solovay (2000, p. 323) constructed an optimal prefix-free universal Turing machine for which ZFC, assuming it is arithmetically sound, cannot prove the value of any individual bit of its halting probability. Arithmetic soundness means that every arithmetical statement proved by the theory is true.

Chaitin constants Omega_U are perhaps the most obvious specific example of noncomputable numbers. They are also known to be transcendental.

For the register machine U and its canonical compressed encoding defined by Calude et al. (2002, §5), let a(n) count the halting programs of total length 7n bits, including both program and data encoded in 7-bit characters. Programs with run-time errors are excluded. The initial counts, starting at n=1, are 1, 1, 3, 8, 50, 311, 1012, 4382, 19164, 99785, 515279, 2559837, ... (OEIS A400878; Calude et al. 2002, Table 1). These counts are specific to this machine and encoding.

This encoding uses an alphabet of 49 characters, each occupying 7 bits, including each data bit. The encoded machine does not satisfy the constant additive program-length overhead required above. Denote its halting probability by H_U.

Grouping the halting programs by total length gives the weighted sum

 H_U=sum_(n=1)^infty(a(n))/(128^n).
(2)

The sum through n=12 is a lower bound for H_U. Calude et al. (2002, §7) bounded the contribution of longer programs to establish the first 64 bits of H_U, whose binary expansion and decimal expansion begin

H_U=0.0000001000000100000110..._2
(3)
=0.00787499699...
(4)

(OEIS A079365 and A100264).

Medallion presented to Gregory Chaitin for his 60th birthday by Stephen Wolfram

Calude and Dinneen (2007) subsequently computed the first 43 and 40 bits of the halting probabilities for two data encodings of a prefix-free Turing machine that is optimal in program length, using data in base 16 and 2, respectively, as

Omega_(U2)=0.0001000000010000101001110111000011111010..._2
(5)
Omega_(U16)=0.0001000000010000101001110111000100000101110..._2.
(6)

The binary result is engraved on the medallion presented to Gregory Chaitin for his 60th birthday by Stephen Wolfram, illustrated above.


See also

Algorithmic Randomness, Computable Number, Halting Probability, Halting Problem, Register Machine, Turing Machine, Universal Turing Machine

Explore with Wolfram|Alpha

References

Borwein, J. and Bailey, D. Mathematics by Experiment: Plausible Reasoning in the 21st Century. Wellesley, MA: A K Peters, p. 145, 2003.Calude, C. S. and Dinneen, M. J. "Exact Approximations of Omega Numbers." Int. J. Bifur. Chaos 17, 1937-1954, 2007. https://doi.org/10.1142/S0218127407018130.Calude, C. S.; Dinneen, M. J.; and Shu, C.-K. "Computing a Glimpse of Randomness." Exper. Math. 11, 361-370, 2002. https://doi.org/10.1080/10586458.2002.10504481.Chaitin, G. Meta Math!: The Quest for Omega. New York: Pantheon Books, 2005.Chaitin, G. J. "A Theory of Program Size Formally Identical to Information Theory." J. Assoc. Comput. Mach. 22, 329-340, 1975. https://doi.org/10.1145/321892.321894.Chaitin, G. J. "How Much Information Can There Be in a Real Number?" Int. J. Bifur. Chaos 17, 1933-1935, 2007.Finch, S. R. "Chaitin's Constant." §1.11 in Mathematical Constants. Cambridge, England: Cambridge University Press, pp. 81-83, 2003.Gardner, M. "The Random Number Omega Bids Fair to Hold the Mysteries of the Universe." Sci. Amer. 241, 20-34, Nov. 1979.Gardner, M. "Chaitin's Omega." Ch. 21 in Fractal Music, Hypercards, and More Mathematical Recreations from Scientific American Magazine. New York: W. H. Freeman, pp. 307-319, 1992.Kobayashi, K. "Sigma_n^0-Complete Properties of Programs and Martin-Löf Randomness." Inform. Process. Lett. 46, 37-42, 1993. https://doi.org/10.1016/0020-0190(93)90194-E.Sloane, N. J. A. Sequences A079365, A100264, and A400878 in "The On-Line Encyclopedia of Integer Sequences."Solovay, R. M. "A Version of Omega for Which ZFC Cannot Predict a Single Bit." In Finite Versus Infinite: Contributions to an Eternal Dilemma (Ed. C. S. Calude and G. Păun). London, England: Springer-Verlag, pp. 323-334, 2000. https://doi.org/10.1007/978-1-4471-0751-4_21.

Referenced on Wolfram|Alpha

Chaitin's Constant

Cite this as:

Weisstein, Eric W. "Chaitin's Constant." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/ChaitinsConstant.html

Subject classifications