A question on mean value inequality

It is known that mean value inequality is very useful. It is:

For any $0 \le a_i (i=1,2,\dots,n)$, $$a_1 a_2\dots a_n\le (\frac{a_1+a_2+\dots + a_n}{n})^n \tag1$$

My question is: how many ways by which the mean value inequality can be proved?

-
@Paul - set $\lambda_i = 1/n$ in the proofs that SYZ's comment links to in order to make the connection – Hans Engler May 31 '13 at 1:08
The inequality don't need to prove. The proof strategy is enough for me. Of course, more details proof are welcome. – Paul May 31 '13 at 1:12
@Paul I saw the tag Integral-inequality, and though I might have misunderstood you post. But I do believe that you are referring to the Arithmetic Mean-Geometric Mean inequality, in which case you can still refer to this AopsWiki page, scroll down and the section below details 2 proofs of non-weighted AM-GM (the one you detailed above). – A Nonny May 31 '13 at 1:17
Wikipedia article gives several proofs of AM-GM nequality. – Martin Sleziak May 31 '13 at 3:52

One method, which is often employed to prove AM-GM, is Cauchy induction. Such proof is sketched in this answer, this answer, this answer, Wikipedia and in many other places.

• Cauchy induction at AoPS
• Proof of AM-GM using this type of induction at Wikipedia. They call this technique forward-backward-induction.
• One section of Pete L. Clark's notes on induction is devoted to this type of induction. He calls it upward-downward induction.
• Perhaps also this question
-