2

Fundamentals

Two beautiful expressions that relate $\pi$ to its convergents are Dalzell integral

$$\frac{22}{7}-\pi=\int_0^1\frac{x^4(1-x)^4}{1+x^2}dx$$

(see Why do we need an integral to prove that $\frac{22}{7} > \pi$?)

and the following equivalent form of Lehmer series $$\pi-3=\sum_{k=1}^\infty \frac{24}{(4 k+1) (4 k+2) (4 k+4)}$$

(see http://matwbn.icm.edu.pl/ksiazki/aa/aa27/aa27121.pdf, page 139)

These are direct proofs that $\pi>3$ and $\frac{22}{7}>\pi$, because of the positiveness of the closed form under the summation and the integral, respectively.

Symmetry

Evaluating Lehmer series leads to the integral $$\pi-3=\int_0^1 \frac{4x^4(1-x)(2+x)}{(1+x)(1+x^2)}dx$$

Another relationship is given by the integral $$\pi-3=\int_{0}^{1} \frac{2 x (1-x)^2}{1+x^2} dx$$

and the series $$\pi-3=\sum_{k=0}^\infty \frac{24}{(4k+2)(4k+3)(4k+5)(4k+6)}$$

A series to prove $\frac{22}{7}-\pi>0$ is given by $$\sum_{k=1}^\infty \frac{240}{(4k+1)(4k+2)(4k+3)(4k+5)(4k+6)(4k+7)}=\frac{22}{7}-\pi$$

which generalizes to $$\sum_{k=n}^\infty \frac{240}{(4k+1)(4k+2)(4k+3)(4k+5)(4k+6)(4k+7)}=\int_0^1 \frac{x^{4n}(1-x)^4}{1+x^2}dx$$

With this expression approximating fractions from the integrals in the RHS may be computed sequentially by adding the next term in the series.

For $n=0$ we have $$ \begin{align} \frac{10}{3}-\pi &= \sum_{k=0}^\infty \frac{240}{(4k+1)(4k+2)(4k+3)(4k+5)(4k+6)(4k+7)} \\ &= \int_0^1 \frac{(1-x)^4}{1+x^2}dx \\ \end{align}$$

and the same difference may be obtained from a series with three factors in the denominator and its corresponding integral

$$\begin{align} \frac{10}{3}-\pi &= \sum_{k=0}^\infty \frac{24}{(4k+4)(4k+6)(4k+7)} \\ &= \int_{0}^{1} \frac{4x^3(1-x)(1+2x)}{(1+x)(1+x^2)}dx \\ \end{align}$$

More convergents

For the third convergent, we have Lucas integral and a series $$\begin{align} \pi-\frac{333}{106} &= \int_0^1 \frac{x^5(1-x)^6(197+462x^2)}{530(1+x^2)}dx \\ &= \frac{48}{371} \sum_{k=0}^\infty \frac{118720 k^2+762311 k+1409424}{(4 k+9) (4 k+11) (4 k+13) (4 k+15) (4 k+17) (4 k+19) (4 k+21) (4 k+23)} \\ \end{align}$$

(see https://math.stackexchange.com/a/1593090/134791)

Finally, Lucas integral for the fourth convergent is $$\frac{355}{113}-\pi=\int_{0}^{1} \frac{x^8(1-x)^8(25+816x^2)}{3164(1+x^2)}dx$$ (see http://educ.jmu.edu/~lucassk/Papers/more%20on%20pi.pdf)

Q Is there a series for $\frac{355}{113}-\pi$?

  • You seem to be really fascinated by this kind of problems. I enjoy to learn all of that from you. Thanks. – Claude Leibovici Feb 14 '16 at 07:38
  • 1
    In what way do the others series correspond to their integrals? – Gerry Myerson Feb 14 '16 at 08:51
  • @ClaudeLeibovici Thank you! They are fascinating and an answer leads to another question... – Jaume Oliver Lafont Feb 14 '16 at 08:57
  • @GerryMyerson They evaluate to the same difference, I do not derive them from each other. – Jaume Oliver Lafont Feb 14 '16 at 08:58
  • Speaking of symmetry in your nice series, factors in the numerators - they are definitely internally symmetrically balanced; for 22/7 one: 3+17=5+15=7+13=9+11=20 for 333/106 one: 9+23=11+21=13+19=15+17=32 – Alex Feb 14 '16 at 14:17
  • @Alex Those denominators are symmetrical by construction, maybe other combinations should be investigated. In contrast, Lehmer series seems like it has a (4k+3) missing, but I would rather have that numerator for the others... Here is how (4k+3) disappears or divides itself into (4k+2) and (4k+4) http://math.stackexchange.com/a/1639961/134791 In the other series, there are constant numerators after solving the system but before merging: http://math.stackexchange.com/questions/1652812/a-series-to-prove-frac227-pi0 maybe that is all that can be expected. – Jaume Oliver Lafont Feb 14 '16 at 14:40
  • 1
    I have no proof but, $\displaystyle \dfrac{355}{113}-\pi=\sum_{k=0}^{+\infty}\dfrac{1489440+10321440k}{1243(4k+13)(4k+15)(4k+17)(4k+19)(4k+21)(4k+23)(4k+25)(4k+27)}$ – FDP Feb 14 '16 at 20:53
  • @FDP Please write that gem as an answer so you get all the upvotes you deserve. Here is the proof, with the common factor in the numerator out of the sum. Thank you very much! http://www.wolframalpha.com/input/?i=(480%2F1243)sum(k%3D0..inf,(3103%2B21503k)%2F((4k%2B13)(4k%2B15)(4k%2B17)(4k%2B19)(4k%2B21)(4k%2B23)(4k%2B25)(4k%2B27))) – Jaume Oliver Lafont Feb 14 '16 at 21:02
  • In that case, Jaume, you are not using the word, "corresponding", the way the rest of the world uses it. – Gerry Myerson Feb 14 '16 at 22:04
  • @GerryMyerson Now there is a true correspondence, I think: http://math.stackexchange.com/a/1657416/134791 Your comment made me think a lot, thank you! – Jaume Oliver Lafont Feb 17 '16 at 20:42

1 Answers1

3

I have no proof of it, but,

$\displaystyle \dfrac{355}{113}-\pi=\sum_{k=0}^{+\infty}\dfrac{1489440+10321440k}{1243(4k+13)(4‌​k+15)(4k+17)(4k+19)(4k+21)(4k+23)(4k+25)(4k+27)}$

I got this "equality" (Wolfram Alpha confirms it) playing around with lindep function of PARI GP (LLL stuff)

FDP
  • 13,647