corollary of Kummer’s theorem


As shown in Kummer’s theorem, the power of a prime numberMathworldPlanetmath p dividing (nm),n≥m∈ℕ, was the total number of carries when adding m and n-m in base p. We’ll give a recurrence relationMathworldPlanetmath for the carry indicator.

Given integers n≥m≥0 and a prime number p, let ni,mi,ri be the i-th digit of n,m, and r:=n-m, respectively.

Define c-1=0, and

ci={1if mi+ri≥p,0otherwise

for each i≥0 up to the number of digits of n.

For each i≥0 we have

ni=mi+ri+ci-1-p.ci.

Starting with the i-th digit of n, we multiply with increasing powers of p to get

∑k=idnk⁢pk-i=(∑k=idpk-i⁢(mk+rk))+∑k=id(pk-1-(i-1)⁢ck-1-pk-(i-1)⁢ck).

The last sum in the above equation leaves only the values for indices i and d, and we get

⌊npi⌋=⌊mpi⌋+⌊rpi⌋+ci-1 (1)

for all i≥0.

Title corollary of Kummer’s theorem
Canonical name CorollaryOfKummersTheorem
Date of creation 2013-03-22 13:23:07
Last modified on 2013-03-22 13:23:07
Owner Thomas Heye (1234)
Last modified by Thomas Heye (1234)
Numerical id 7
Author Thomas Heye (1234)
Entry type Corollary
Classification msc 11A63