question

Suppose that p, 2p-1, 3p-2 are all primes, with p>3 . Prove that p(2p-1)(3p-2) is a Carmichael number. Find the smallest Carmichael number of this form.

Thank you very much.

Printable View

- Jan 14th 2007, 09:05 PMJenny20carmichael numbers
question

Suppose that p, 2p-1, 3p-2 are all primes, with p>3 . Prove that p(2p-1)(3p-2) is a Carmichael number. Find the smallest Carmichael number of this form.

Thank you very much. - Jan 15th 2007, 07:01 AMThePerfectHacker
First I need to explain what a Carmichael number is.

Any prime satisfiers this (Fermat's little theorem):

$\displaystyle a^{p}\equiv a (\mbox{mod } p)$

For any integer $\displaystyle a$.

However, the converse fails.

For example,

$\displaystyle 2^{341}\equiv 2 (\mbox{mod }341)$.

This is an example of a*pseudoprime*(to base two) for which the converse fails.

A pseudoprime to all base is call a*Carmichaeal*number.

----

Since $\displaystyle p,2p-1,3p-2$ are all primes we have.

Let $\displaystyle \gcd(a,p)=\gcd(a,2p-1)=\gcd(a,3p-2)=1$.

Then, Fermat's Elegant theorem (different version than above)

$\displaystyle a^{p-1} \equiv 1 (\mbox{mod }p)$

$\displaystyle a^{2p-2}\equiv 1 (\mbox{mod }2p-1)$

$\displaystyle a^{3p-3}\equiv 1 (\mbox{mod }3p-2)$

Thus,

$\displaystyle (a^{p-1})^{6p^2-(p-1)}\equiv 1(\mbox{mod }p)$

$\displaystyle (a^{2p-2})^{3p^2-(p-1)/2}\equiv 1(\mbox{mod }2p-1)$

$\displaystyle (a^{3p-3})^{2p^2-(p-1)/3}\equiv 1(\mbox{mod }3p-2)$

**Important**Note $\displaystyle (p-1)/2$ is an integer because primes are odd (usually) and we require for $\displaystyle (p-1)/3$ to be an integer we need that $\displaystyle p\equiv 1(\mbox{mod }3)$.

Now, because of relative primeness in the moduli we have,

$\displaystyle a^{6p^3-7p^2+2p-1}\equiv 1(\mbox{mod }p(2p-1)(3p-2)=6p^3-7p^2+2p)$

Thus,

$\displaystyle a^{6p^3-7p^2+2p}\equiv a(\mbox{mod }6p^3-7p^2+2p)$.

Thus, if $\displaystyle p\equiv 1(\mbox{mod }3)$.

And, $\displaystyle p,2p-1,3p-2$ are primes.

Then,

$\displaystyle p(2p-1)(3p-2)$ is a Carmichael number.

---

I am supprised, where did you get this theorem from? First time I hear of it. It is really interesting. - May 9th 2009, 06:52 PMMedia_ManRewriteQuote:

Suppose that p, 2p-1, 3p-2 are all primes, with p>3 . Prove that p(2p-1)(3p-2) is a Carmichael number.

Carmichael Number -- from Wolfram MathWorld