Jump to content

Irrationality measure

From Wikipedia, the free encyclopedia
Rational approximations to the square root of 2.

In mathematics, an irrationality measure of a real number x is a measure of how closely it can be approximated by a rational number.

If a function f(t,λ), defined for t,λ>0, takes positive real values and is strictly decreasing in both variables, consider the following inequality:

0<|x−pq|<f(q,λ)

for a given real number x∈ℝ and rational numbers pq with p∈ℤ,q∈ℤ+. For x∈ℝ, define R as the set of all λ∈ℝ+ for which the set {pq∈ℚ:|x−pq|<f(q,λ)} is finite. Then λ(x)=inf⁡R is called an irrationality measure of x with regard to f. If the set R is empty, x is said to have infinite irrationality measure λ(x)=∞.

By definition, there is a sequence (εn)n∈ℕ converging to 0 such that for every n, the inequality

0<|x−pq|<f(q,λ(x)+εn)

has at most finitely many solutions pq. By the nondecreasing property of f in the second variable, we deduce that the inequality

0<|x−pq|<f(q,λ(x)+ε)

has at most only finitely many solutions pq for all ε>0.[1]

Irrationality exponent

[edit]

The irrationality exponent or Liouville–Roth irrationality measure is given by setting f(q,μ)=q−μ,[1] a definition adapting the one of Liouville numbers — the irrationality exponent μ(x) is defined for real numbers x to be the supremum of the set of μ such that 0<|x−pq|<1qμ is satisfied by an infinite number of coprime integer pairs (p,q) with q>0.[2][3]: 246 

For any value n<μ(x), the infinite set of all rationals p/q satisfying the above inequality yields good approximations of x. Conversely, if n>μ(x), then there are at most finitely many coprime (p,q) with q>0 that satisfy the inequality.

For example, whenever a rational approximation pq≈x with p,q∈ℕ yields n+1 exact decimal digits, then

110n≥|x−pq|≥1qμ(x)+ε

for any ε>0, except for at most a finite number of "lucky" pairs (p,q).

A number x∈ℝ with irrationality exponent μ(x)≤2 is called a diophantine number,[4] while numbers with μ(x)=∞ are called Liouville numbers.

Corollaries

[edit]

Rational numbers have irrationality exponent 1, while (as a consequence of Dirichlet's approximation theorem) every irrational number has irrationality exponent at least 2.

On the other hand, an application of Borel-Cantelli lemma shows that almost all numbers, including all algebraic irrational numbers, have an irrationality exponent exactly equal to 2.[3]: 246 

It is μ(x)=μ(rx+s) for real numbers x and rational numbers r≠0 and s. If for some x we have μ(x)≤μ, then it follows μ(x1/2)≤2μ.[5]: 368 

For a real number x given by its simple continued fraction expansion x=[a0;a1,a2,...] with convergents pi/qi it holds:[1]

μ(x)=1+lim supn→∞ln⁡qn+1ln⁡qn=2+lim supn→∞ln⁡an+1ln⁡qn.

If we have lim supn→∞1nln⁡|qn|≤σ and limn→∞1nln⁡|qnx−pn|=−τ for some positive real numbers σ,τ, then we can establish an upper bound for the irrationality exponent of x by:[6][7]

μ(x)≤1+στ

Known bounds

[edit]

For most transcendental numbers, the exact value of their irrationality exponent is not known.[5] Below is a table of known upper and lower bounds.

Number x Irrationality exponent μ(x) Notes
Lower bound Upper bound
Rational number p/q with p∈ℤ,q∈ℤ+ 1 Every rational number p/q has an irrationality exponent of exactly 1.
Irrational algebraic number α 2 By Roth's theorem the irrationality exponent of any irrational algebraic number is exactly 2. Examples include square roots and the golden ratio φ.
e2/k,k∈ℤ+ 2 If the elements an of the simple continued fraction expansion of an irrational number x are bounded above an<P(n) by an arbitrary polynomial P, then its irrationality exponent is μ(x)=2.

Examples include numbers whose continued fractions behave predictably such as

e=[2;1,2,1,1,4,1,1,6,1,...] and I0(2)/I1(2)=[1;2,3,4,5,6,7,8,9,10,...].

tan⁡(1/k),k∈ℤ+ 2
tanh⁡(1/k),k∈ℤ+ 2
S(b) with b≥2 2 S(b):=∑k=0∞b−2kwith b∈ℤ, has continued fraction terms which do not exceed a fixed constant.[8][9]
T(b) with b≥2[10] 2 T(b):=∑k=0∞tkb−k where tk is the Thue–Morse sequence and b∈ℤ. See Prouhet-Thue-Morse constant.
ln⁡(2)[11][12] 2 3.57455... There are other numbers of the form ln⁡(a/b) for which bounds on their irrationality exponents are known.[13][14][15]
ln⁡(3)[11][16] 2 5.11620...
5ln⁡(3/2)[17] 2 3.43506... There are many other numbers of the form 2k+1ln⁡(2k+1+12k+1−1) for which bounds on their irrationality exponents are known.[17] This is the case for k=12.
π/3[18][19] 2 4.60105... There are many other numbers of the form 2k−1arctan⁡(2k−1k−1) for which bounds on their irrationality exponents are known.[18] This is the case for k=2.
π[11][20] 2 7.10320... It has been proven that π's irrationality exponent is at most 5/2 if the Flint Hills series ∑n=1∞csc2nn3 (where n is in radians) converges,[21][22] and that if the irrationality exponent is at least 5/2 it diverges.[23] Conversely,[23] (theorem 2.5) proved that the Flint Hills series converges if π's irrationality exponent is ≤5/2, completing the biconditional (unless μ(π)=5/2). In 2026, OpenAI released an unverified proof claiming that μ(π)=2.[24]
π2[11][25] 2 5.09541... π2 and ζ(2) are linearly dependent over ℚ. (ζ(2)=π26), also see the Basel problem.
arctan⁡(1/2)[26] 2 9.27204... There are many other numbers of the form arctan⁡(1/k) for which bounds on their irrationality exponents are known.[27][28]
arctan⁡(1/3)[29] 2 5.94202...
Apéry's constant ζ(3)[11] 2 5.51389...
Γ(1/4)[30] 2 10143
Cahen's constant C[31] 3
Champernowne constants Cb in base b≥2[32] b Examples include C10=0.1234567891011...=[0;8,9,1,149083,1,...]
Liouville numbers L ∞ The Liouville numbers are precisely those numbers having infinite irrationality exponent.[3]: 248 

Irrationality base

[edit]

The irrationality base or Sondow irrationality measure is obtained by setting f(q,β)=β−q.[1][6] It is a weaker irrationality measure, being able to distinguish how well different Liouville numbers can be approximated, but yielding β(x)=1 for all other real numbers:

Let x be an irrational number. If there exist real numbers β≥1 with the property that for any ε>0, there is a positive integer q(ε) such that

|x−pq|>1(β+ε)q

for all integers p,q with q≥q(ε) then the least such β is called the irrationality base of x and is represented as β(x).

If no such β exists, then β(x)=∞ and x is called a super Liouville number.

If a real number x is given by its simple continued fraction expansion x=[a0;a1,a2,...] with convergents pi/qi then it holds:

β(x)=lim supn→∞ln⁡qn+1qn=lim supn→∞ln⁡an+1qn.[1]

Examples

[edit]

Any real number x with finite irrationality exponent μ(x)<∞ has irrationality base β(x)=1, while any number with irrationality base β(x)>1 has irrationality exponent μ(x)=∞ and is a Liouville number.

The number L=[1;2,22,222,...] has irrationality exponent μ(L)=∞ and irrationality base β(L)=1.

The numbers τa=∑n=0∞1na=1+1a+1aa+1aaa+1aaaa+... (na represents tetration, a=2,3,4...) have irrationality base β(τa)=a.

The number S=1+121+1421+18421+1168421+132168421+… has irrationality base β(S)=∞, hence it is a super Liouville number.

Although it is not known whether or not eπ is a Liouville number,[33]: 20  it is known that β(eπ)=1.[5]: 371 

Other irrationality measures

[edit]

Markov constant

[edit]

Setting f(q,M)=(Mq2)−1 gives a stronger irrationality measure: the Markov constant M(x). For an irrational number x∈ℝ∖ℚ it is the factor by which Dirichlet's approximation theorem can be improved for x. Namely if c<M(x) is a positive real number, then the inequality

0<|x−pq|<1cq2

has infinitely many solutions pq∈ℚ. If c>M(x) there are at most finitely many solutions.

Dirichlet's approximation theorem implies M(x)≥1 and Hurwitz's theorem gives M(x)≥5 both for irrational x.[34]

This is in fact the best general lower bound since the golden ratio gives M(φ)=5. It is also M(2)=22.

Given x=[a0;a1,a2,...] by its simple continued fraction expansion, one may obtain:[35]

M(x)=lim supn→∞([an+1;an+2,an+3,...]+[0;an,an−1,...,a2,a1]).

Bounds for the Markov constant of x=[a0;a1,a2,...] can also be given by p2+4≤M(x)<p+2 with p=lim supn→∞an.[36] This implies that M(x)=∞ if and only if (ak) is not bounded. The numbers with finite M(x) are called badly approximable. In particular every quadratic irrational number x is badly approximable.

Any number with μ(x)>2 or β(x)>1 has an unbounded simple continued fraction and hence M(x)=∞. A further consequence is M(e)=∞.

For rational numbers r it may be defined M(r)=0.

Other results

[edit]

The values M(e)=∞ and μ(e)=2 imply that the inequality 0<|e−pq|<1cq2 has for all c∈ℝ+ infinitely many solutions pq∈ℚ while the inequality 0<|e−pq|<1q2+ε has for all ε∈ℝ+ only at most finitely many solutions pq∈ℚ. This gives rise to the question what the best upper bound is. The answer is given by:[37]

0<|e−pq|<cln⁡ln⁡qq2ln⁡q

which is satisfied by infinitely many pq∈ℚ for c>12 but not for c<12.

This makes the number e alongside the rationals and quadratic irrationals an exception to the fact that for almost all real numbers x∈ℝ the inequality below has infinitely many solutions pq∈ℚ:[5] (see Khinchin's theorem)

0<|x−pq|<1q2ln⁡q

Mahler's generalization

[edit]

Kurt Mahler extended the concept of an irrationality measure and defined a so-called transcendence measure, drawing on the idea of a Liouville number and partitioning the transcendental numbers into three distinct classes.[3]

Mahler's irrationality measure

[edit]

Instead of taking for a given real number x the difference |x−p/q| with p/q∈ℚ, one may instead focus on term |qx−p|=|L(x)| with p,q∈ℤ and L∈ℤ[x] with deg⁡L=1. Consider the following inequality:

0<|qx−p|≤max⁡(|p|,|q|)−ω with p,q∈ℤ and ω∈ℝ0+.

Define R as the set of all ω∈ℝ0+ for which infinitely many solutions p,q∈ℤ exist, such that the inequality is satisfied. Then ω1(x)=sup⁡M is Mahler's irrationality measure. It gives ω1(p/q)=0 for rational numbers, ω1(α)=1 for algebraic irrational numbers and in general ω1(x)=μ(x)−1, where μ(x) denotes the irrationality exponent.

Transcendence measure

[edit]

Mahler's irrationality measure can be generalized as follows:[2][3] Take P to be a polynomial with deg⁡P≤n∈ℤ+ and integer coefficients ai∈ℤ. Then define a height function H(P)=max⁡(|a0|,|a1|,...,|an|) and consider for complex numbers z the inequality:

0<|P(z)|≤H(P)−ω with ω∈ℝ0+.

Set R to be the set of all ω∈ℝ0+ for which infinitely many such polynomials exist, that keep the inequality satisfied. Further define ωn(z)=sup⁡R for all n∈ℤ+ with ω1(z) being the above irrationality measure, ω2(z) being a non-quadraticity measure, etc.

Then Mahler's transcendence measure is given by:

ω(z)=lim supn→∞ωn(z).

The transcendental numbers can now be divided into the following three classes:

If for all n∈ℤ+ the value of ωn(z) is finite and ω(z) is finite as well, z is called an S-number (of type ω(z)).

If for all n∈ℤ+ the value of ωn(z) is finite but ω(z) is infinite, z is called an T-number.

If there exists a smallest positive integer N such that for all n≥N the ωn(z) are infinite, z is called an U-number (of degree N).

The number z is algebraic (and called an A-number) if and only if ω(z)=0.

Almost all numbers are S-numbers. In fact, almost all real numbers give ω(x)=1 while almost all complex numbers give ω(z)=12.[38]: 86  The number e is an S-number with ω(e)=1. The number π is either an S- or T-number.[38]: 86  The U-numbers are a set of measure 0 but still uncountable.[39] They contain the Liouville numbers which are exactly the U-numbers of degree one.

Linear independence measure

[edit]

Another generalization of Mahler's irrationality measure gives a linear independence measure.[2][13] For real numbers x1,...,xn∈ℝ consider the inequality

0<|c1x1+...+cnxn|≤max⁡(|c1|,...,|cn|)−ν with c1,...,cn∈ℤ and ν∈ℝ0+.

Define R as the set of all ν∈ℝ0+ for which infinitely many solutions c1,...cn∈ℤ exist, such that the inequality is satisfied. Then ν(x1,...,xn)=sup⁡R is the linear independence measure.

If the x1,...,xn are linearly dependent over ℚ then ν(x1,...,xn)=0.

If 1,x1,...,xn are linearly independent algebraic numbers over ℚ then ν(1,x1,...,xn)≤n.[33]

It is further ν(1,x)=ω1(x)=μ(x)−1.

Other generalizations

[edit]

Koksma’s generalization

[edit]

Jurjen Koksma in 1939 proposed another generalization, similar to that of Mahler, based on approximations of complex numbers by algebraic numbers.[3][38]

For a given complex number z consider algebraic numbers α of degree at most n. Define a height function H(α)=H(P), where P is the characteristic polynomial of α and consider the inequality:

0<|z−α|≤H(α)−ω∗−1 with ω∗∈ℝ0+.

Set R to be the set of all ω∗∈ℝ0+ for which infinitely many such algebraic numbers α exist, that keep the inequality satisfied. Further define ωn∗(z)=sup⁡R for all n∈ℤ+ with ω1∗(z) being an irrationality measure, ω2∗(z) being a non-quadraticity measure,[17] etc.

Then Koksma's transcendence measure is given by:

ω∗(z)=lim supn→∞ωn∗(z).

The complex numbers can now once again be partitioned into four classes A*, S*, T* and U*. However it turns out that these classes are equivalent to the ones given by Mahler in the sense that they produce exactly the same partition.[38]: 87 

Simultaneous approximation of real numbers

[edit]

Given a real number x∈ℝ, an irrationality measure of x quantifies how well it can be approximated by rational numbers pq with denominator q∈ℤ+. If x=α is taken to be an algebraic number that is also irrational one may obtain that the inequality

0<|α−pq|<1qμ

has only at most finitely many solutions pq∈ℚ for μ>2. This is known as Roth's theorem.

This can be generalized: Given a set of real numbers x1,...,xn∈ℝ one can quantify how well they can be approximated simultaneously by rational numbers p1q,...,pnq with the same denominator q∈ℤ+. If the xi=αi are taken to be algebraic numbers, such that 1,α1,...,αn are linearly independent over the rational numbers ℚ it follows that the inequalities

0<|αi−piq|<1qμ,∀i∈{1,...,n}

have only at most finitely many solutions (p1q,...,pnq)∈ℚn for μ>1+1n. This result is due to Wolfgang M. Schmidt.[40][41]

See also

[edit]

References

[edit]
  1. 1 2 3 4 5 Sondow, Jonathan (2004). "Irrationality Measures, Irrationality Bases, and a Theorem of Jarnik". arXiv:math/0406300.
  2. 1 2 3 Parshin, A. N.; Shafarevich, I. R. (2013-03-09). Number Theory IV: Transcendental Numbers. Springer Science & Business Media. ISBN 978-3-662-03644-0.
  3. 1 2 3 4 5 6 Bugeaud, Yann (2012). Distribution modulo one and Diophantine approximation. Cambridge Tracts in Mathematics. Vol. 193. Cambridge: Cambridge University Press. doi:10.1017/CBO9781139017732. ISBN 978-0-521-11169-0. MR 2953186. Zbl 1260.11001.
  4. ↑ Tao, Terence (2009). "245B, Notes 9: The Baire category theorem and its Banach space consequences". What's new. Retrieved 2024-09-08.
  5. 1 2 3 4 Borwein, Jonathan M. (1987). Pi and the AGM: A Study in Analytic Number Theory and Computational Complexity. Wiley.
  6. 1 2 Sondow, Jonathan (2003-07-23). "An irrationality measure for Liouville numbers and conditional measures for Euler's constant". arXiv:math/0307308.
  7. ↑ Chudnovsky, G. V. (1982). "Hermite-padé approximations to exponential functions and elementary estimates of the measure of irrationality of π". In Chudnovsky, David V.; Chudnovsky, Gregory V. (eds.). The Riemann Problem, Complete Integrability and Arithmetic Applications. Lecture Notes in Mathematics. Vol. 925. Berlin, Heidelberg: Springer. pp. 299–322. doi:10.1007/BFb0093516. ISBN 978-3-540-39152-4.
  8. ↑ Shallit, Jeffrey (1979-05-01). "Simple continued fractions for some irrational numbers". Journal of Number Theory. 11 (2): 209–217. doi:10.1016/0022-314X(79)90040-4. ISSN 0022-314X.
  9. ↑ Shallit, J. O (1982-04-01). "Simple continued fractions for some irrational numbers, II". Journal of Number Theory. 14 (2): 228–231. doi:10.1016/0022-314X(82)90047-6. ISSN 0022-314X.
  10. ↑ Bugeaud, Yann (2011). "On the rational approximation to the Thue–Morse–Mahler numbers". Annales de l'Institut Fourier. 61 (5): 2065–2076. doi:10.5802/aif.2666. ISSN 1777-5310.
  11. 1 2 3 4 5 Weisstein, Eric W. "Irrationality Measure". mathworld.wolfram.com. Retrieved 2020-10-14.
  12. ↑ Nesterenko, Yu. V. (2010-10-01). "On the irrationality exponent of the number ln 2". Mathematical Notes. 88 (3): 530–543. doi:10.1134/S0001434610090257. ISSN 1573-8876. S2CID 120685006.
  13. 1 2 Wu, Qiang (2003). "On the Linear Independence Measure of Logarithms of Rational Numbers". Mathematics of Computation. 72 (242): 901–911. doi:10.1090/S0025-5718-02-01442-4. ISSN 0025-5718. JSTOR 4099938.
  14. ↑ Bouchelaghem, Abderraouf; He, Yuxin; Li, Yuanhang; Wu, Qiang (2024-03-01). "On the linear independence measures of logarithms of rational numbers. II". J. Korean Math. Soc. 61 (2): 293–307. doi:10.4134/JKMS.j230133.
  15. ↑ Sal’nikova, E. S. (2008-04-01). "Diophantine approximations of log 2 and other logarithms". Mathematical Notes. 83 (3): 389–398. doi:10.1134/S0001434608030097. ISSN 1573-8876.
  16. ↑ "Symmetrized polynomials in a problem of estimating of the irrationality measure of number ln 3". www.mathnet.ru. Retrieved 2020-10-14.
  17. 1 2 3 Polyanskii, Alexandr (2015-01-27). "On the irrationality measure of certain numbers". arXiv:1501.06752 [math.NT].
  18. 1 2 Polyanskii, A. A. (2018-03-01). "On the Irrationality Measures of Certain Numbers. II". Mathematical Notes. 103 (3): 626–634. doi:10.1134/S0001434618030306. ISSN 1573-8876. S2CID 125251520.
  19. ↑ Androsenko, V. A. (2015). "Irrationality measure of the number \frac{\pi}{\sqrt{3}}". Izvestiya: Mathematics. 79 (1): 1–17. doi:10.1070/im2015v079n01abeh002731. ISSN 1064-5632. S2CID 123775303.
  20. ↑ Zeilberger, Doron; Zudilin, Wadim (2020-01-07). "The irrationality measure of π is at most 7.103205334137...". Moscow Journal of Combinatorics and Number Theory. 9 (4): 407–419. arXiv:1912.06345. doi:10.2140/moscow.2020.9.407. S2CID 209370638.
  21. ↑ Alekseyev, Max A. (2011). "On convergence of the Flint Hills series". arXiv:1104.5100 [math.CA].
  22. ↑ Weisstein, Eric W. "Flint Hills Series". MathWorld.
  23. 1 2 Meiburg, Alex (2022). "Bounds on Irrationality Measures and the Flint-Hills Series". arXiv:2208.13356 [math.NT].
  24. ↑ OpenAI (2026-09-24). "The irrationality exponent of π is 2" (PDF). GitHub. Retrieved 2026-10-07.
  25. ↑ Zudilin, Wadim (2014-06-01). "Two hypergeometric tales and a new irrationality measure of ζ(2)". Annales mathématiques du Québec. 38 (1): 101–117. arXiv:1310.1526. doi:10.1007/s40316-014-0016-0. ISSN 2195-4763. S2CID 119154009.
  26. ↑ Bashmakova, M. G.; Salikhov, V. Kh. (2019). "Об оценке меры иррациональности arctg 1/2". Чебышевский сборник. 20 (4 (72)): 58–68. ISSN 2226-8383.
  27. ↑ Tomashevskaya, E. B. "On the irrationality measure of the number log 5+pi/2 and some other numbers". www.mathnet.ru. Retrieved 2020-10-14.
  28. ↑ Salikhov, Vladislav K.; Bashmakova, Mariya G. (2022). "On rational approximations for some values of arctan(s/r) for natural s and r, s". Moscow Journal of Combinatorics and Number Theory. 11 (2): 181–188. doi:10.2140/moscow.2022.11.181. ISSN 2220-5438.
  29. ↑ Salikhov, V. Kh.; Bashmakova, M. G. (2020-12-01). "On Irrationality Measure of Some Values of $\operatorname{arctg} \frac{1}{n}$". Russian Mathematics. 64 (12): 29–37. doi:10.3103/S1066369X2012004X. ISSN 1934-810X.
  30. ↑ Bruiltet, Sylvain (2002). "D'une mesure d'approximation simultanee a une mesure d'irrationalite : le cas de Γ(1/4) et Γ(1/3)". Acta Arithmetica. 104 (3): 243–281. doi:10.4064/aa104-3-3. ISSN 0065-1036.
  31. ↑ Duverney, Daniel; Shiokawa, Iekata (2020-01-01). "Irrationality exponents of numbers related with Cahen's constant". Monatshefte für Mathematik. 191 (1): 53–76. doi:10.1007/s00605-019-01335-0. ISSN 1436-5081.
  32. ↑ Amou, Masaaki (1991-02-01). "Approximation to certain transcendental decimal fractions by algebraic numbers". Journal of Number Theory. 37 (2): 231–241. doi:10.1016/S0022-314X(05)80039-3. ISSN 0022-314X.
  33. 1 2 Waldschmidt, Michel (2004-01-24). "Open Diophantine Problems". arXiv:math/0312440.
  34. ↑ Hurwitz, A. (1891). "Ueber die angenäherte Darstellung der Irrationalzahlen durch rationale Brüche (On the approximate representation of irrational numbers by rational fractions)". Mathematische Annalen (in German). 39 (2): 279–284. doi:10.1007/BF01206656. JFM 23.0222.02. S2CID 119535189.
  35. ↑ LeVeque, William (1977). Fundamentals of Number Theory. Addison-Wesley Publishing Company, Inc. pp. 251–254. ISBN 0-201-04287-8.
  36. ↑ Hancl, Jaroslav (January 2016). "Second basic theorem of Hurwitz". Lithuanian Mathematical Journal. 56: 72–76. doi:10.1007/s10986-016-9305-4. S2CID 124639896.
  37. ↑ Davis, C. S. (1978). "Rational approximations to e". Journal of the Australian Mathematical Society. 25 (4): 497–502. doi:10.1017/S1446788700021480. ISSN 1446-8107.
  38. 1 2 3 4 Baker, Alan (1979). Transcendental number theory (Repr. with additional material ed.). Cambridge: Cambridge Univ. Pr. ISBN 978-0-521-20461-3.
  39. ↑ Burger, Edward B.; Tubbs, Robert (2004-07-28). Making Transcendence Transparent: An Intuitive Approach to Classical Transcendental Number Theory. Springer Science & Business Media. ISBN 978-0-387-21444-3.
  40. ↑ Schmidt, Wolfgang M. (1972). "Norm Form Equations". Annals of Mathematics. 96 (3): 526–551. doi:10.2307/1970824. ISSN 0003-486X. JSTOR 1970824.
  41. ↑ Schmidt, Wolfgang M. (1996). Diophantine approximation. Lecture notes in mathematics. Berlin; New York: Springer. ISBN 978-3-540-09762-4.