Let k,p be positive integers. Is there a closed form for the sums Sum_(i=0)^p (k i) (k+p-i p-i), or Sum_(i=0)^p (k-1 i) (k+p-i p-i)?

Audrey Arnold

Audrey Arnold

Answered question

2022-11-17

Let k , p be positive integers. Is there a closed form for the sums
i = 0 p ( k i ) ( k + p i p i ) , or
i = 0 p ( k 1 i ) ( k + p i p i ) ?

Answer & Explanation

Nkgopotsev1g

Nkgopotsev1g

Beginner2022-11-18Added 15 answers

Derived the following simple inequality (confirmed numerically): for any 0 < s < 1 / 2,
i = 0 p ( k i ) ( k + p i p i ) 1 ( 1 2 s ) k ( 1 s s ) p .
Given 0 < s < 1 / 2, let X be a binomial ( k , s ) random variable, and Y a binomial ( k + p X , t ) random variable, where t = s / ( 1 s ) ( ( 0 , 1 ) ). Then, by the law of total probability,
P ( X + Y = p ) = i = 0 p P ( X + Y = p | X = i ) P ( X = i ) = i = 0 p P ( Y = p i | X = i ) P ( X = i ) .
Noting that
P ( X = i ) = ( k i ) s i ( 1 s ) k i
and
P ( Y = p i | X = i ) = ( k + p i p i ) t p i ( 1 t ) k ,
and using
s ( 1 s ) t = 1 ,
we get
P ( X + Y = p ) = ( 1 s ) k ( 1 t ) k t p i = 0 p ( k i ) ( k + p i p i ) .
Finally, from P ( X + Y = p ) 1 and
( 1 s ) k ( 1 t ) k t p = ( 1 2 s ) k ( s 1 s ) p ,
it follows that
i = 0 p ( k i ) ( k + p i p i ) 1 ( 1 2 s ) k ( 1 s s ) p .

Do you have a similar question?

Recalculate according to your conditions!

Ask your question.
Get an expert answer.

Let our experts help you. Answer in as fast as 15 minutes.

Didn't find what you were looking for?