|
hbghlyj
Posted at 2023-2-19 20:52:05
Last edited by hbghlyj at 2023-2-20 01:31:00Asymptote

en.wikipedia.org/wiki/Gaussian_integral
generalisation of Gaussian integral
multidimensional Gaussian integral
下面是Feynman integration的方法 (相关帖子)
As before, we simplify matters by computing the right half of the integral, $\int_0^{\infty} e^{-x^2} d x$. Set
$$
I(t):=\int_0^{\infty} \frac{e^{-x^2}}{1+(x / t)^2} d x .
$$
This is defined for all $t>0$, and we're hoping to determine $I(\infty)$. We simplify this integral by making the natural substitution:
$$
I(t)=t \int_0^{\infty} \frac{e^{-t^2 x^2}}{1+x^2} d x
$$
Right away this tells us new information about $I(t)$, for example, that
$$\tag2\label2
\lim _{t \rightarrow 0} \frac{I(t)}{t}=\frac{\pi}{2} .
$$
(In fact, the left hand side of this is $I^{\prime}(0)$, but we won't make use of this below.) To apply Feynman's trick we want to find $I^{\prime}(t)$, but if we do this naively the outcome will be ugly. To simplify matters, we recall our warm-up example $G(t)$, in which case differentiating simplified the integral because the denominator cancelled! Inspired by this, we instead work with the function
$$
e^{-t^2} I(t)=t \int_0^{\infty} \frac{e^{-t^2\left(1+x^2\right)}}{1+x^2} d x
$$
Finally, to avoid the product rule we divide both sides by $t$. Having thus renormalized $I$, we differentiate:
$$
\frac{d}{d t}\left(t^{-1} e^{-t^2} I(t)\right)=\int_0^{\infty}-2 t e^{-t^2\left(1+x^2\right)} d x=-2 e^{-t^2} \int_0^{\infty} e^{-u^2} d u=-2 e^{-t^2} I(\infty)
$$
Now, somewhat ridiculously, we undo what we just did by integrating both sides:
$$
\underbrace{\int_0^{\infty} \frac{d}{d t}\left(t^{-1} e^{-t^2} I(t)\right) d t}_{=-\lim _{t \rightarrow 0} \frac{I(t)}{t}}=\underbrace{\int_0^{\infty}-2 e^{-t^2} I(\infty) d t}_{=-2 I(\infty)^2}
$$
Since we computed the left hand side in \eqref{2}, we deduce $I(\infty)^2=\frac{\pi}{4}$. It follows that $I(\infty)=\frac{\sqrt{\pi}}{2}$, whence
$$
\int_{-\infty}^{\infty} e^{-x^2} d x=\sqrt{\pi}
$$ |
|