MathDL - The MAA Mathematical Sciences Digital Library
Search

Search Loci: Convergence:

Keyword

  Advanced Search
Random Quotation

Einstein, Albert (1879-1955)

Everything should be made as simple as possible, but not simpler.

Reader's Digest. Oct. 1977.

See more quotations

The Mathematical Association of America
The National Science Digital Library Project
The National Science Foundation
Register Sign In

Loci: Convergence

Sums of Powers of Positive Integers

by Janet Beery (University of Redlands)

Solutions to Exercises 9-12

Exercise 9. 

Diagram after al-Haytham

Figure 19.  Each side of the equation $$(4 + 1)\sum_{i = 1}^4 {i^2 }  = \sum_{i = 1}^4 {i^3 }  + \sum_{p = 1}^4 {\sum_{i = 1}^p {i^2 } }$$ is the area of the rectangle.

 

Exercise 10.   Let k = 1 in Figure 7.  Then each side of the equation $$(n + 1)\sum_{i = 1}^n i = \sum_{i = 1}^n {i^2 }  + \sum_{p = 1}^n {\sum_{i = 1}^p i }$$

is the area of the rectangle.  Letting k = 1 in the equation $$(n + 1)\sum_{i = 1}^n {i^k}  =  \sum_{i = 1}^n {i^{k + 1}}  + \sum_{p = 1}^n {\sum_{i = 1}^p {i^k }}$$

gives the equation $$(n + 1)\sum_{i = 1}^n i = \sum_{i = 1}^n {i^2 }  + \sum_{p = 1}^n {\sum_{i = 1}^p i }$$ or

\(\displaystyle{(n + 1)(1 + 2 + 3 +  \cdots  + n) =}\)  $$(1^2  + 2^2  + 3^2  +  \cdots  + n^2 ) + \sum_{p = 1}^n {(1 + 2 + 3 +  \cdots  + p)}$$ or

 \(\displaystyle{1^2  + 2^2  + 3^2  +  \cdots  + n^2  =}\) $$(n + 1)(1 + 2 + 3 +  \cdots  + n) - \sum_{p = 1}^n {(1 + 2 + 3 +  \cdots  + p)}$$ or $$\eqalign{1^2  + 2^2  + 3^2  +  \cdots  + n^2 & = (n + 1)\left( {{{n(n + 1)} \over 2}} \right) - \sum_{p = 1}^n {{{p(p + 1)} \over 2}}\cr   &= (n + 1)\left( {{{n(n + 1)} \over 2}} \right) - {1 \over 2}\sum_{p = 1}^n {p^2 }  - {1 \over 2}\sum_{p = 1}^n p \cr &= (n + 1)\left( {{{n(n + 1)} \over 2}} \right) - {1 \over 2}\left( {1^2  + 2^2  + 3^2  +  \cdots  + n^2 } \right)}$$

\(\displaystyle{- {1 \over 2}\left( {1 + 2 + 3 +  \cdots  + n} \right)}\)

or $$\eqalign{{3 \over 2}\left( {1^2  + 2^2  + 3^2  +  \cdots  + n^2 } \right) &= (n + 1)\left( {{{n(n + 1)} \over 2}} \right) - {1 \over 2}{{n(n + 1)} \over 2}\cr &= \left( {n + {1 \over 2}} \right)\left( {{{n(n + 1)} \over 2}} \right)}$$ or $$1^2  + 2^2  + 3^2  +  \cdots  + n^2  = {{n(n + 1)(2n + 1)} \over 6}.$$

 

Exercise 11.  Let k = 2 in Figure 7.  Then each side of the equation $$(n + 1)\sum_{i = 1}^n i^2 = \sum_{i = 1}^n {i^3 }  + \sum_{p = 1}^n {\sum_{i = 1}^p i^2}$$

is the area of the rectangle.  Letting k = 2 in the equation $$(n + 1)\sum_{i = 1}^n {i^k}  =  \sum_{i = 1}^n {i^{k + 1}}  + \sum_{p = 1}^n {\sum_{i = 1}^p {i^k }}$$

gives the equation $$(n + 1)\sum_{i = 1}^n i^2 = \sum_{i = 1}^n {i^3}  + \sum_{p = 1}^n {\sum_{i = 1}^p i^2 }$$ or

\(\displaystyle{(n + 1)(1^2  + 2^2  + 3^2  +  \cdots  + n^2 ) =}\) $$(1^3  + 2^3  + 3^3  +  \cdots  + n^3 ) + \sum_{p = 1}^n {(1^2  + 2^2  + 3^2  +  \cdots  + p^2 )}$$ or

 \(\displaystyle{1^3  + 2^3  + 3^3  +  \cdots  + n^3  =}\) $$(n + 1)(1^2  + 2^2  + 3^2  +  \cdots  + n^2 ) - \sum_{p = 1}^n {(1^2  + 2^2  + 3^2  +  \cdots  + p^2 )}.$$

If we now apply the formula for the sum of the squares, the equation becomes

$$1^3  + 2^3  + 3^3  +  \cdots  + n^3 = (n + 1)\left( {{{n(n + 1)(2n + 1)} \over 6}} \right) - \sum_{p = 1}^n {{{p(p + 1)(2p + 1)} \over 6}}$$ $$= (n + 1)\left( {{{n(n + 1)(2n + 1)} \over 6}} \right) - {1 \over 3}\sum_{p = 1}^n {p^3 }  - {1 \over 2}\sum_{p = 1}^n {p^2 }  - {1 \over 6}\sum_{p = 1}^n p$$

$$ = (n + 1)\left( {{{n(n + 1)(2n + 1)} \over 6}} \right) - {1 \over 3}\left( {1^3  + 2^3  + 3^3  +  \cdots  + n^3 } \right)$$

\(\displaystyle{- {1 \over 2}\left( {1^2  + 2^2  + 3^2  +  \cdots  + n^2 } \right) - {1 \over 6}\left( {1 + 2 + 3 +  \cdots  + n} \right)}.\)

Collecting sums of cubes and applying the formula for the sum of the squares again, we have

\(\displaystyle{{4 \over 3}\left( 1^3  + 2^3  + 3^3  +  \cdots  + n^3 \right) =}\) $$(n + 1)\left( {n(n + 1)(2n + 1) \over 6} \right) - {1 \over 2}{{n(n + 1)(2n + 1)} \over 6} - {1 \over 6}{{n(n + 1)} \over 2}$$ or $$\eqalign{1^3  + 2^3  + 3^3  +  \cdots  + n^3  &= {3 \over 4}\left( {{{n(n + 1)} \over {12}}\left( {2(n + 1)(2n + 1) - (2n + 1) - 1} \right)} \right)\cr  &= {1 \over 4}\left( {{{n(n + 1)} \over 4}\left( {(2n + 1)^2  - 1} \right)} \right)\cr &= {1 \over 4}\left( {{{n(n + 1)} \over 4}\left( {4n^2  + 4n} \right)} \right)}$$ or $$1^3  + 2^3  + 3^3  +  \cdots  + n^3 = \left( n(n + 1) \over 2 \right)^2.$$

 

Exercise 12.

$$\eqalign{\left({n \over 5} + {1 \over 5} \right)n\left( n + {1 \over 2} \right)\left( (n + 1)n - {1 \over 3} \right) &={1 \over 5}(n + 1)n\left({1 \over 2} \right)(2n + 1)\left((n + 1)n - {1 \over 3} \right)\cr &={n(n + 1)(2n + 1) \over 6}\left({3 \over 5}(n + 1)n - {1 \over 5} \right).}$$

Pages: | 1 |  2 |  3 |  4 |  5 |  6 |  7 |  8 |  9 |  10 |  11 |  12 |  13 |  14 |  15 |  16 |  17 |  18 |  19 |  20 |  21 | 

Beery, Janet, "Sums of Powers of Positive Integers," Loci (February 2009), DOI: 10.4169/loci003284



Discuss this article

start a new discussion thread  | show all 3 threads about this article

thread #1:

Karaji's solution

by Hasan Unal (posted: 01/09/2010 )

I just wonder if the Karaji Generalized his method of solutions? Is there any information about that. For instance did he found sum of fifth powers?

Reply:

Al-Karaji and higher powers by Janet Beery (posted: 01/20/2010 )
My understanding is that there is no evidence that al-Karaji derived a "formula" for the sum of the fourth, fifth, or higher powers. His justification for his formula for the sum of the cubes was ingenious but al-Haytham's idea seems more readily generalizable.

add your reply

thread #2:

On the symmetry of the resulting polynomials

by Christopher Yeleighton (posted: 05/01/2011 )

== Conjecture == It would be interesting to know that * sums of odd powers can be factored as a polynomial on (n ⋅ (n + 1) / 2) and is therefore symmetric with the center at (−½); and that * sums of even powers can be factored to a function like above multiplied by (2 ⋅ n + 1) and therefore antisymmetric with the center at (−½). It would be interesting to know whether any rule concerning the polynomials thus obtained.

Replies:

Re: On the symmetry of the resulting polynomials by Rick Mabry (posted: 12/30/2011 )
Christopher, see the following article, especially section 3 (Faulhaber Polynomials). Although my browsers don't let me view the code you pasted, I think the article confirms some of what you're getting at (some of which might be gleaned from page 8 of the article we're viewing). A. F. Beardon, Sums of Powers of Integers, American Mathematical Monthly, Vol. 103, No. 3 (Mar., 1996), pp. 201-213.

More about symmetry: by Janet Beery (posted: 01/03/2013 )
The most succinct, beautiful, and perhaps even historically plausible explanation I've seen for the symmetry of the Faulhaber polynomials about -1/2 was in a recent article by Reuben Hersh in the College Math Journal 43:4 (Sept. 2012), pp. 322-4. His approach? Extend the polynomials to negative values of n.

add your reply

MathDL Homepage MathDL Homepage National Science Digital Library The Mathematical Association of America