Logarithmic mean

From Wikipedia, the free encyclopedia

In mathematics, the logarithmic mean is a function of two numbers which is equal to their difference divided by the logarithm of their quotient. In symbols:

\begin{matrix} M_{\mbox{lm}}(x,y) &= \lim_{(\xi,\eta)\to(x,y)} \frac{\eta - \xi}{\ln \eta - \ln \xi} \\ &= \begin{cases} x & \mbox{if }x=y \\ \frac{y - x}{\ln y - \ln x} & \mbox{else} \end{cases} \end{matrix}

for the positive numbers x,y. This measure is useful in engineering problems involving heat and mass transfer.

Contents

[edit] Inequalities

The logarithmic mean of two numbers is smaller than the arithmetic mean but larger than the geometric mean (unless the numbers are the same of course, in which case all three means are equal to the numbers):

\forall x>0\ \forall y>0\ x\ne y\Rightarrow \sqrt{x\cdot y} < \frac{y - x}{\ln y - \ln x} < \frac{x+y}{2}

[edit] Derivation of the mean

[edit] Mean value theorem of differential calculus

From the mean value theorem

\exists \xi\in[x,y]\ f'(\xi) = \frac{f(x)-f(y)}{x-y}

the logarithmic mean is obtained as the value of ξ by substituting ln for f

\frac{1}{\xi} = \frac{\ln x - \ln y}{x-y}

and solving for ξ.

\xi = \frac{x-y}{\ln x - \ln y}

[edit] Integration

The logarithmic mean can also be interpreted as the area under an exponential curve.

L(x,y) = \int_0^1 x^{1-t}\cdot y^t\ \mathrm{d}t

(Check \begin{array}{rcl} \int_0^1 x^{1-t}\cdot y^t\ \mathrm{d}t &=& \int_0^1 \left(\frac{y}{x}\right)^t\cdot x\ \mathrm{d}t \\ &=& x\cdot \int_0^1 \exp\left(t\cdot\ln \frac{y}{x}\right) \mathrm{d}t \\ &=& \frac{x}{\ln \frac{y}{x}} \cdot \left[ \exp\left(t\cdot\ln \frac{y}{x}\right) \right]_{t=0}^{1} \\ &=& \frac{x}{\ln \frac{y}{x}} \cdot \left(\frac{y}{x}-1\right) \end{array})

The area interpretation allows to easily derive basic properties of the logarithmic mean. Since the exponential function is monotonic the integral over an interval of length 1 is bounded by x and y. The Homogenity of the integral operator is transferred to the mean operator, that is L(c\cdot x, c\cdot y) = c\cdot L(x,y).

[edit] Generalization

[edit] Mean value theorem of differential calculus

You can generalize the mean to n + 1 variables by considering the mean value theorem for divided differences for the nth derivative of the logarithm. You obtain

L_{\mathrm{MV}}(x_0,\dots,x_n) = \sqrt[-n]{(-1)^{(n+1)}\cdot n \cdot \ln[x_0,\dots,x_n]}

where \ln[x_0,\dots,x_n] denotes a divided difference of the logarithm.

For n = 2 this leads to

L_{\mathrm{MV}}(x,y,z) = \sqrt{\frac{(x-y)\cdot(y-z)\cdot(z-x)}{2\cdot((y-z)\cdot\ln x + (z-x)\cdot\ln y + (x-y)\cdot\ln z)}}.

[edit] Integral

The integral interpretation can also be generalized to more variables, but it leads to a different result. Given the simplex S with S = \{(\alpha_0,\dots,\alpha_n) : \alpha_0+\dots+\alpha_n=1\ \land\ \alpha_0\ge0\ \land\ \dots\ \land\ \alpha_n\ge0\} and an appropriate measure dα which assigns the simplex a volume of 1, we obtain

L_{\mathrm{I}}(x_0,\dots,x_n) = \int_S x_0^{\alpha_0}\cdot\dots\cdot x_n^{\alpha_n}\ \mathrm{d}\alpha

This can be simplified using divided differences of the exponential function to

L_{\mathrm{I}}(x_0,\dots,x_n) = n!\cdot\exp[\ln x_0, \dots, \ln x_n].

Example n = 2

L_{\mathrm{I}}(x,y,z) = -2\cdot\frac{x\cdot(\ln y-\ln z) + y\cdot(\ln z-\ln x) + z\cdot(\ln x-\ln y)}{(\ln x-\ln y)\cdot(\ln y-\ln z)\cdot(\ln z-\ln x)}.

[edit] See also

[edit] References