Monday, 2 May 2011

nt.number theory - How to put "x^a => 1/(a+1) x^(a+1)" and "x^-1 => log(x)" together

First of all, I'll repeat what I said in the comments: writing fracxh1h=fracehlogx1h makes it fairly clear that as hto0, this expression tends to logx, and setting h=a+1 this is precisely the desired result.



I should mention that this result is implicit in a certain fact well-known to people who do competition math, which is as follows. Given non-negative real numbers x1,...xn, let Ap(x1,...xn)=sqrt[p]fracxp1+...+xpnn denote the p-power mean for pneq0. For p=0, define A0(x1,...xn)=sqrt[n]x1...xn (the geometric mean, and also the limit as pto0 of the above).



Theorem (Power Mean Inequality): If pleq, then ApleAq.



If you like fancy keywords, then I will bring to your attention that as ptoinfty the p-power mean approaches textmax(x1,...xn), which one can think of as the "low-temperature limit" of ordinary addition becoming tropical addition. Then pto0 can be thought of as the "high-temperature limit," in which ordinary addition becomes multiplication instead. Somebody who knows more statistical mechanics than I do (that is, any) can probably tell you the physical significance of this.

No comments:

Post a Comment