Proof that 22/7 exceeds π explained

Proofs of the mathematical result that the rational number is greater than (pi) date back to antiquity. One of these proofs, more recently developed but requiring only elementary techniques from calculus, has attracted attention in modern mathematics due to its mathematical elegance and its connections to the theory of Diophantine approximations. Stephen Lucas calls this proof "one of the more beautiful results related to approximating ".Julian Havil ends a discussion of continued fraction approximations of with the result, describing it as "impossible to resist mentioning" in that context.

The purpose of the proof is not primarily to convince its readers that is indeed bigger than ; systematic methods of computing the value of exist. If one knows that is approximately 3.14159, then it trivially follows that  < , which is approximately 3.142857. But it takes much less work to show that  <  by the method used in this proof than to show that is approximately 3.14159.

Background

is a widely used Diophantine approximation of . It is a convergent in the simple continued fraction expansion of . It is greater than, as can be readily seen in the decimal expansions of these values:

\begin{align} (22)/(7) & = 3 . \overline{ 142857 }, \\ \pi & = 3 . 141 59265 \ldots \end{align}

The approximation has been known since antiquity. Archimedes wrote the first known proof that is an overestimate in the 3rd century BCE, although he may not have been the first to use that approximation. His proof proceeds by showing that is greater than the ratio of the perimeter of a regular polygon with 96 sides to the diameter of a circle it circumscribes.

Proof

The proof can be expressed very succinctly:

0 < \int ^{1} _{0} (x ^{4\left(1-x\right)} ^{4})/(1+ x ^{2}) dx = (22)/(7) - \pi.

Therefore,  > .

The evaluation of this integral was the first problem in the 1968 Putnam Competition.It is easier than most Putnam Competition problems, but the competition often features seemingly obscure problems that turn out to refer to something very familiar. This integral has also been used in the entrance examinations for the Indian Institutes of Technology.^[1]

Details of evaluation of the integral

That the integral is positive follows from the fact that the integrand is non-negative; the denominator is positive and the numerator is a product of nonnegative numbers. One can also easily check that the integrand is strictly positive for at least one point in the range of integration, say at . Since the integrand is continuous at that point and nonnegative elsewhere, the integral from 0 to 1 must be strictly positive.

It remains to show that the integral in fact evaluates to the desired quantity:

\begin{align} 0 & < \int ^{1 (x ^{4\left(1-x\right)} ^{4})/(1+ x ^{2})} _{0} dx \\[ 8 pt] & = \int ^{1 (x ^{4-4x} ^{5+6x} ^{6-4x} ^{7+x} ^{8})/(1+ x ^{2})} _{0} dx & expansionoftermsinthenumerator \\[ 8 pt] & = \int ^{1} _{0} \left(x ^{6-4x} ^{5+5x} ^{4-4x} ^{2+4- (4)/(1+ x ^{2})} \right) dx & usingpolynomiallongdivision & \\[ 8 pt] & = \left.\left( (x ^{7})/(- 7) (2 x ^{6})/(3) + x ^{5-} (4 x ^{3})/(3) +4 x -4 \arctan{x}\right)\right| ^{1} _{0} & definiteintegration \\[ 6 pt] & = (1)/(- 7) (2)/(+1- 3) (4)/(3) +4- \pi & with \arctan( 1 ) = (\pi)/(4) and \arctan( 0 ) = 0 \\[ 8 pt] & = (22)/(7) - \pi. & addition \end{align}

(See polynomial long division.)

Quick upper and lower bounds

In, it is pointed out that if 1 is substituted for in the denominator, one gets a lower bound on the integral, and if 0 is substituted for in the denominator, one gets an upper bound:^[2]

(1)/(1260) = \int ^{1 (x ^{4} \left( 1- x\right) ^{4})/(2)} _{0} dx < \int ^{1 (x ^{4} \left( 1- x\right) ^{4})/(1+ x ^{2})} _{0} dx < \int ^{1 (x ^{4} \left( 1- x\right) ^{4})/(1)} _{0} dx = { 1 \over 630 }.

Thus we have

(22)/(7) - (1)/(630) < \pi < (22)/(7) - (1)/(1260) ,

hence 3.1412 < < 3.1421 in decimal expansion. The bounds deviate by less than 0.015% from . See also .^[3]

Proof that 355/113 exceeds

As discussed in, the well-known Diophantine approximation and far better upper estimate for follows from the relation

0 <\int ^{1 (x ^{8\left(1-x\right)} ^{8\left(25+816x} ^{2\right)})/(3164 \left( 1+ x ^{2\right)})} _{0} dx = (355)/(113) - \pi. (355)/(113) = 3 . 14159292 \ldots,

where the first six digits after the decimal point agree with those of . Substituting 1 for in the denominator, we get the lower bound

\int ^{1 (x ^{8\left(1-x\right)} ^{8\left(25+816x} ^{2\right)})/(6328)} _{0} dx = (911)/(5261111856) = 0 . 000000173 \ldots,

substituting 0 for in the denominator, we get twice this value as an upper bound, hence

(355)/(- 113) (911)/(<\pi< 2630555928) (355)/(- 113) (911)/(5261111856) .

In decimal expansion, this means, where the bold digits of the lower and upper bound are those of .

Extensions

The above ideas can be generalized to get better approximations of ; see also and (in both references, however, no calculations are given). For explicit calculations, consider, for every integer,

(1 { 2)/(^{2n-1}) }\int_0^1 x^(1-x)^\,dx<\frac1\int_0^1\frac\,dx<\frac1\int_0^1 x^(1-x)^\,dx,

where the middle integral evaluates to

\begin{align} (1 { 2)/(^{2n-2}) } & \int_0^1\frac\,dx\\[6pt]= & \sum_^\frac+(-1)^n\left(\pi-4\sum_^\frac\right)\end

involving . The last sum also appears in Leibniz' formula for . The correction term and error bound is given by

\begin{align} (1 { 2)/(^{2n-1}) }\int_0^1 x^(1-x)^\,dx&=\frac\\[6pt]&\sim\frac,\end

where the approximation (the tilde means that the quotient of both sides tends to one for large) of the central binomial coefficient follows from Stirling's formula and shows the fast convergence of the integrals to .

Calculation of these integrals: For all integers and we have

\begin{align} x ^{k(1-x)} ^{\ell&=(1-2x+x} ^{2)x} ^{k(1-x)} ^{\ell-2} \\[ 6 pt] & = ( 1+ x ^{2)x} ^{k(1-x)} ^{\ell-2} -2 x ^{k+1} ( 1- x) ^{\ell-2} . \end{align}

Applying this formula recursively times yields

x ^{4n} ( 1- x) ^{4n} = \left( 1+ x ^{2 n -1} _{j =0} ( -2 ) ^{jx} ^{4n+j} ( 1- x) ^{4n-2(j+1)} + ( -2 ) ^{2n} x ^{6n} .

Furthermore,

\begin{align} x ^{6n} - ( -1 ) ^{3n} & = \sum ^{3 n} _{j =1} ( -1 ) ^{3n-j} x ^{2j} - \sum ^{3 n -1} _{j =0} ( -1 ) ^{3n-j} x ^{2j} \\[ 6 pt] & = \sum ^{3 n -1} _{j =0} \left(( -1 ) ^{3n-(j+1)} x ^{2(j+1)} - ( -1 ) ^{3n-j} x ^{2j} \right)\\[ 6 pt] & =- ( 1+ x ^{3 n -1} _{j =0} ( -1 ) ^{3n-j} x ^{2j} , \end{align}

where the first equality holds, because the terms for cancel, and the second equality arises from the index shift in the first sum.

Application of these two results gives

\begin{align} (x ^{4n} ( 1- x) ^{4n})/(2 ^{2n-2} ( 1+ x ^{2)}) = \sum ^{2 n -1} _{j =0} & (( -1 ) ^{j})/(2 ^{2n-j-2}) x ^{4n+j} ( 1- x) ^{4n-2j-2} \\[ 6 pt] & {} -4 \sum ^{3 n -1} _{j =0} ( -1 ) ^{3n-j} x ^{2j} + ( -1 ) ^{3n} (4 { 1+ x)/(^{2}. (1) \end{align}})

For integers, using integration by parts times, we obtain

\begin{align} \int ^{1 x} _{0} ^{k(1-x)} ^{\elldx & = (\ell{k +1 }\int)/(_{0})} ^{1x} ^{k+1} ( 1- x) ^{\ell-1} dx\\[ 6 pt] &\vdots\\[ 6 pt] & = (\ell{k +1 })/(… (\ell -1)/(k +2)) (1 {k + \ell}\int)/(_{0}) ^{1x} ^{k+\ell} dx\\[ 6 pt] & = (1)/((k + \ell +1 )\binom{k + \ell) {k}}. ( 2 ) \end{align}

Setting, we obtain

\int ^{1} _{0} x ^{4n} ( 1- x) ^{4n} dx = (1)/(( 8 n +1 )\binom{ 8 n) { 4 n}}.

Integrating equation (1) from 0 to 1 using equation (2) and, we get the claimed equation involving .

The results for are given above. For we get

(14 \int)/(_{0}) ^{1 (x ^{8(1-x)} ^{8})/(1+ x ^{2})} dx = \pi - (47171)/(15015)

and

(18 \int)/(_{0}) ^{1} x ^{8(1-x)} ^{8 dx = (1 { 1750320 },)/()}

hence, where the bold digits of the lower and upper bound are those of . Similarly for,

(1 { 16 }\int)/(_{0}) ^{1 (x ^{12} \left( 1- x\right) ^{12})/(1+ x ^{2})} dx = (431302721)/(137287920) - \pi

with correction term and error bound

(1 { 32 }\int)/(_{0}) ^{1} x ^{12} ( 1- x) ^{12} dx = (1 { 2163324800 },)/()

hence . The next step for is

(1 { 64 }\int)/(_{0}) ^{1 (x ^{16} ( 1- x) ^{16})/(1+ x ^{2})} dx = \pi - (741269838109)/(235953517800)

with

(1 { 128 }\int)/(_{0}) ^{1} x ^{16} ( 1- x) ^{16} dx = (1 { 2538963567360 },)/()

which gives .

See also

• Approximations of
• Chronology of computation of
• Lindemann–Weierstrass theorem (proof that is transcendental)
• List of topics related to
• Proof that is irrational

Footnotes

Citations

External links

• The problems of the 1968 Putnam competition, with this proof listed as question A1.

Notes and References

1. https://drive.google.com/file/d/0BxIR7Kj71SMKUmxnUkdhdWJ5NWM/view 2010 IIT Joint Entrance Exam
2. .
3. .