In the derivation of the integral representation of the beta function, why is the upper limit 1 after changing variables?

224 Views Asked by At

In the derivation of the beta function the following proof is given:

$\begin{align} \Gamma(x)\Gamma(y) &= \int_{u=0}^\infty\ e^{-u} u^{x-1}\,du \cdot\int_{v=0}^\infty\ e^{-v} v^{y-1}\,dv \\[6pt] &=\int_{v=0}^\infty\int_{u=0}^\infty\ e^{-u-v} u^{x-1}v^{y-1}\,du \,dv. \end{align}$

Changing variables by u=zt and v=z(1-t) shows that this is

$\begin{align} \Gamma(x)\Gamma(y) &= \int_{z=0}^\infty\int_{t=0}^\color{red}{1} e^{-z} (zt)^{x-1}(z(1-t))^{y-1}\big|J(z,t)\big|\,dt \,dz \tag{1}\\[6pt] &= \int_{z=0}^\infty\int_{t=0}^\color{red}{1} e^{-z} (zt)^{x-1}(z(1-t))^{y-1}z\,dt \,dz \\[6pt] &= \int_{z=0}^\infty e^{-z}z^{x+y-1} \,dz\cdot\int_{t=0}^\color{red}{1}t^{x-1}(1-t)^{y-1}\,dt\\ &=\Gamma(x+y)\,\beta(x,y), \end{align}$

How was the upper limit of integration calculated after the change of variables. I thought I understood the Fubini-Tonnelli theorem but I don't understand how the value of 1 was calculated. Why is there no $\frac{1}{z}$ dependence in the upper limit of (1)?


In a single dimension its easy to get the limits 0,1 using u=1/x-1 but in the two dimensional plane I'm confused by this substitution.

1

There are 1 best solutions below

0
On BEST ANSWER

The easiest way to see what the limits of integration should be is to write out the inverse transformations. In this case, we have $$ u + v = zt + z(1 - t) = z \quad \Rightarrow \quad z = u + v $$ and $$\frac{v}{u} = \frac{1 - t}{t} = \frac{1}{t} - 1 \quad \Rightarrow \quad t = \frac{1}{v/u + 1}. $$

We then need to look at the range of values these quantities can take on when $u \in (0, \infty)$ and $v \in (0, \infty)$. For $z$, $0 < u < \infty$ and $0 < v < \infty$ imply that $0 < z < \infty$. The values of $t$ are a little more subtle. The ratio $v/u$ also satisfies $0 < v/u < \infty$; in the limit $v/u \to 0$, we have $t \to 1$, while in the limit $v/u \to \infty$ we have $t \to 0$. Since this is an invertible one-to-one coordinate transformation, we can conclude that the limits of integration under this change of variables are $z \in (0, \infty)$ and $t \in (0,1)$.