The Student Room Group

approximation of sum of integers to the power of 4

Also, I know how to prove sum of squares but how would you do power of 4?

Thanks!
Reply 1
Original post by sangyoonknow
Also, I know how to prove sum of squares but how would you do power of 4?

Thanks!


Also, this makes no sense. :tongue:
Reply 2
Original post by sangyoonknow
Also, I know how to prove sum of squares but how would you do power of 4?

Thanks!


Can you expand on what you're trying to do?
Reply 3
Original post by davros
Can you expand on what you're trying to do?



Original post by BabyMaths
Also, this makes no sense. :tongue:



Original post by sangyoonknow
Also, I know how to prove sum of squares but how would you do power of 4?

Thanks!



In suspect s(he) means 1^4 + 2^4 + 3^4 + 4^4 + .... + n^4
Reply 4
Original post by TeeEm
In suspect s(he) means 1^4 + 2^4 + 3^4 + 4^4 + .... + n^4


I suspect so too, but the onus is on the OP to tell us :smile:
1^4 + 2^4 + 3^4 + 4^4 + .... + n^4!

How would I approximate the value of it for n=100?
Reply 6
Original post by sangyoonknow
1^4 + 2^4 + 3^4 + 4^4 + .... + n^4!

How would I approximate the value of it for n=100?


Do you want an approximation or the exact value?

There will be a formula for the sum, but I'm not going to do the heavy algebra at this time in the evening :smile:

If you know the methods for summing n^2 and n^3 it's the same in principle, just a little messier!
Original post by sangyoonknow
1^4 + 2^4 + 3^4 + 4^4 + .... + n^4!

How would I approximate the value of it for n=100?

How closely do you want to approximate it? Zero is a pretty good approximation, only out by at most 10^20.

Otherwise, the sum is going to be given by a fifth-power expression, and will be divisible by 16n(n+1)(2n+1)\frac{1}{6}n(n+1)(2n+1) being the corresponding sum of squares; hence it is a quadratic times that thing. (It's a fact, I seem to remember, that the sum of nth powers is a multiple of the sum of n-2th powers.)

That thing is 338350, so we're looking at roughly 1002×338350=3,383,500,000100^2 \times 338350 = 3,383,500,000. I'll knock some sig figs off that because there's no reason at all for the quadratic to have leading coefficient 1; let's go with 1×10101 \times 10^{10} as the possibly-correct order of magnitude.

Alternatively:

i=1ni4=nE(X4)\sum_{i=1}^n i^4 = n \mathbb{E}(X^4) where XX follows the uniform distribution on {1,2,,n}\{1, 2, \dots, n \}. Also i=2n+1i4=nE((X+1)4)\sum_{i=2}^{n+1} i^4 = n \mathbb{E}((X+1)^4), so (n+1)41=nE((X+1)4X4)=nE[4X3+6X2+4X+1](n+1)^4-1 = n \mathbb{E}((X+1)^4 - X^4) = n \mathbb{E}[4X^3+6X^2+4X+1].

That is, n3+4n2+6n+4=4E(X3)+6E(X2)+4E(X)+1n^3+4n^2+6n+4 = 4 \mathbb{E}(X^3) + 6 \mathbb{E}(X^2) + 4 \mathbb{E}(X) + 1. E(X)=n+12\mathbb{E}(X) = \frac{n+1}{2}; E(X2)=12(n+1)(2n+1)\mathbb{E}(X^2) = \frac{1}{2} (n+1)(2n+1) by the usual formula for the sum of squares; hence E(X3)\mathbb{E}(X^3) can be calculated by simply rearranging.
if you consider r=1n(r+1)krk\displaystyle\sum_{r=1}^{n} (r+1)^k - r^k you should be able to derive a formula for general integer values of k in terms of the sums of the previous powers :smile:
(edited 9 years ago)
Original post by Smaug123
How closely do you want to approximate it? Zero is a pretty good approximation, only out by at most 10^20.

Otherwise, the sum is going to be given by a fifth-power expression, and will be divisible by 16n(n+1)(2n+1)\frac{1}{6}n(n+1)(2n+1) being the corresponding sum of squares; hence it is a quadratic times that thing. (It's a fact, I seem to remember, that the sum of nth powers is a multiple of the sum of n-2th powers.)

That thing is 338350, so we're looking at roughly 1002×338350=3,383,500,000100^2 \times 338350 = 3,383,500,000. I'll knock some sig figs off that because there's no reason at all for the quadratic to have leading coefficient 1; let's go with 1×10101 \times 10^{10} as the possibly-correct order of magnitude.

Alternatively:

i=1ni4=nE(X4)\sum_{i=1}^n i^4 = n \mathbb{E}(X^4) where XX follows the uniform distribution on {1,2,,n}\{1, 2, \dots, n \}. Also i=2n+1i4=nE((X+1)4)\sum_{i=2}^{n+1} i^4 = n \mathbb{E}((X+1)^4), so (n+1)41=nE((X+1)4X4)=nE[4X3+6X2+4X+1](n+1)^4-1 = n \mathbb{E}((X+1)^4 - X^4) = n \mathbb{E}[4X^3+6X^2+4X+1].

That is, n3+4n2+6n+4=4E(X3)+6E(X2)+4E(X)+1n^3+4n^2+6n+4 = 4 \mathbb{E}(X^3) + 6 \mathbb{E}(X^2) + 4 \mathbb{E}(X) + 1. E(X)=n+12\mathbb{E}(X) = \frac{n+1}{2}; E(X2)=12(n+1)(2n+1)\mathbb{E}(X^2) = \frac{1}{2} (n+1)(2n+1) by the usual formula for the sum of squares; hence E(X3)\mathbb{E}(X^3) can be calculated by simply rearranging.



Sorry I don't really understand

Otherwise, the sum is going to be given by a fifth-power expression, and will be divisible by being the corresponding sum of squares; hence it is a quadratic times that thing. (It's a fact, I seem to remember, that the sum of nth powers is a multiple of the sum of n-2th powers.)
Original post by Arithmeticae
if you consider r=1n(r+1)krk\displaystyle\sum_{r=1}^{n} (r+1)^k - r^k you should be able to derive a formula for general integer values of k in terms of the sums of the previous powers :smile:



Yes that is the way I figured out how to do! But I was asked to approximate it in my practise interview and couln't think of any way to do it without using the method above.
Reply 11
Original post by sangyoonknow
Yes that is the way I figured out how to do! But I was asked to approximate it in my practise interview and couln't think of any way to do it without using the method above.


sum of n is O(n2)
sum of n2 is O(n3)
sum of n3 is O(n4)
sum of n4 is O(n5)

therefore 1005

bit crude but there are approximations and approximations
Original post by sangyoonknow
Yes that is the way I figured out how to do! But I was asked to approximate it in my practise interview and couln't think of any way to do it without using the method above.


Based on an integral (n+12)55\dfrac{\left(n+\frac{1}{2}\right)^5}{5} is OK?
Original post by TeeEm
sum of n is O(n2)
sum of n2 is O(n3)
sum of n3 is O(n4)
sum of n4 is O(n5)

therefore 1005

bit crude but there are approximations and approximations


But how did you get to that approximation of putting a pwoer up?
Reply 14
Original post by sangyoonknow
But how did you get to that approximation of putting a pwoer up?


look at the formulas of

Sum(n)
Sum(n2)
Sum(n3)
Original post by TeeEm
look at the formulas of

Sum(n)
Sum(n2)
Sum(n3)


ohhhh!!! I feel so stupid! I got it now! thank you so much:smile:

Could you possibly explain to me about:

will be divisible by being the corresponding sum of squares; hence it is a quadratic times that thing. (It's a fact, I seem to remember, that the sum of nth powers is a multiple of the sum of n-2th powers.)
Reply 16
Original post by sangyoonknow
ohhhh!!! I feel so stupid! I got it now! thank you so much:smile:

Could you possibly explain to me about:

will be divisible by being the corresponding sum of squares; hence it is a quadratic times that thing. (It's a fact, I seem to remember, that the sum of nth powers is a multiple of the sum of n-2th powers.)


You are correct, and even more frustrating for me is that I used to remember the proof, but at the moment I do not even remember the quadratic formula.Hopefully somebody else will.:smile:
Original post by sangyoonknow
Sorry I don't really understand

Otherwise, the sum is going to be given by a fifth-power expression, and will be divisible by being the corresponding sum of squares; hence it is a quadratic times that thing. (It's a fact, I seem to remember, that the sum of nth powers is a multiple of the sum of n-2th powers.)

I said the bracketed sentence because I forget the proof, but it is true that the sum of the nth powers is a multiple of the sum of the n-2th powers. Let n=4 to get the first sentence, which implicitly assumes also that the sum of n^k is O(x^{k+1}).

Quick Reply

Latest