查看“︁刘维尔公式”︁的源代码

'''刘维尔公式'''（Liouville's Formula）是一个关于[[多重积分]]和[[欧拉积分]]（<math>\Gamma</math>函数）的公式，其形式如下：
:<math>\int ...\iint_{x_{1},x_{2},...,x_{n}\geqslant0;x_{1}+x_{2}+...+x_{n}\leqslant1}f\left(x_{1}+x_{2}+...+x_{n}\right)x_{1}^{p_{1}-1}x_{2}^{p_{2}-1}...x_{n}^{p_{n}-1}\mathrm{d}x_{1}\mathrm{d}x_{2}...\mathrm{d}x_{n}</math>

<math>=\frac{\Gamma\left(p_{1}\right)\Gamma\left(p_{2}\right)...\Gamma\left(p_{n}\right)}{\Gamma\left(p_{1}+p_{2}+...+p_{n}\right)}\int_{0}^{1}f\left(u\right)u^{p_{1}+p_{2}+...+p_{n}-1}\mathrm{d}u</math>

其中<math>p_{1},p_{2},...,p_{n}>0</math>，<math>f\left(u\right)</math>为[[连续函数]]。<ref name="a1">Б.П.吉米多维奇. 《吉米多维奇数学分析习题集题解》. 济南: 山东科学技术出版社. 2014. ISBN 978-7-5331-5895-8.</ref>

==证明==
用[[数学归纳法]]。
当n=1时，公式显然成立。

当n=2时，公式也成立，即
:<math>\iint_{x_{1},x_{2}\geqslant0;x_{1}+x_{2}\leqslant1}f\left(x_{1}+x_{2}\right)x_{1}^{p_{1}-1}x_{2}^{p_{2}-1}\mathrm{d}x_{1}\mathrm{d}x_{2}=\frac{\Gamma\left(p_{1}\right)\Gamma\left(p_{2}\right)}{\Gamma\left(p_{1}+p_{2}\right)}\int_{0}^{1}f\left(u\right)u^{p_{1}+p_{2}-1}\mathrm{d}u</math>

事实上，令<math>\Omega</math>表示区域：<math>x_{1}\geqslant0,x_{2}\geqslant0,x_{1}+x_{2}\leqslant1</math>，作代换<math>x_{1}=\xi_{1},x_{1}+x_{2}=\xi_{2}</math>，以及<math>t=\frac{\xi_{1}}{\xi_{2}}</math>，则有
:<math>\iint_{x_{1},x_{2}\geqslant0;x_{1}+x_{2}\leqslant1}f\left(x_{1}+x_{2}\right)x_{1}^{p_{1}-1}x_{2}^{p_{2}-1}\mathrm{d}x_{1}\mathrm{d}x_{2}=\int_{0}^{1}f\left(\xi_{2}\right)\mathrm{d}\xi_{2}\int_{0}^{\xi_{2}}\xi_{1}^{p_{1}-1}\left(\xi_{2}-\xi{1}\right)^{p_{2}-1}\mathrm{d}\xi_{1}</math>
:<math>\int_{0}^{1}f\left(\xi_{2}\right)\mathrm{d}\xi_{2}\int_{0}^{1}t^{p_{1}-1}\left(1-t\right)^{p_{2}-1}\xi_{2}^{p_{1}+p_{2}-1}\mathrm{d}t=\frac{\Gamma\left(p_{1}\right)\Gamma\left(p_{2}\right)}{\Gamma\left(p_{1}+p_{2}\right)}\int_{0}^{1}f\left(\xi_{2}\right)\xi_{2}^{p_{1}+p_{2}-1}\mathrm{d}\xi_{2}=\frac{\Gamma\left(p_{1}\right)\Gamma\left(p_{2}\right)}{\Gamma\left(p_{1}+p_{2}\right)}\int_{0}^{1}f\left(u\right)u^{p_{1}+p_{2}-1}\mathrm{d}u</math>

设公式对于n-1成立，今证对于n公式也成立。为此，将公式左端写为
:<math>\int ...\iint_{x_{1},x_{2},...,x_{n-1}\geqslant0;x_{1}+x_{2}+...+x_{n-1}\leqslant1}x_{1}^{p_{1}-1}x_{2}^{p_{2}-1}...x_{n-1}^{p_{n-1}-1}\mathrm{d}x_{1}\mathrm{d}x_{2}...\mathrm{d}x_{n-1}\int_{0}^{1-\left(x_{1}+x_{2}+...+x_{n-1}\right)}f\left(x_{1}+x_{2}+...+x_{n}\right)x_{n}^{p_{n}-1}\mathrm{d}x_{n}</math>

令<math>\psi\left(s\right)=\int_{0}^{1-s}f\left(s+x_{n}\right)x_{n}^{p_{n}-1}\mathrm{d}x_{n}</math>

代入上式，并利用公式对{{mvar|n}}-1成立的假定，得知上式为
:<math>\frac{\Gamma\left(p_{1}\right)\Gamma\left(p_{2}\right)...\Gamma\left(p_{n-1}\right)}{\Gamma\left(p_{1}+p_{2}+...+p_{n-1}\right)}\int_{0}^{1}\psi\left(s\right)s^{p_{1}+p_{2}+...+p_{n-1}-1}\mathrm{d}s</math>

利用上面已证的{{mvar|n}}=2时的公式，于是即得
:<math>\int ...\iint_{x_{1},x_{2},...,x_{n}\geqslant0;x_{1}+x_{2}+...+x_{n}\leqslant1}f\left(x_{1}+x_{2}+...+x_{n}\right)x_{1}^{p_{1}-1}x_{2}^{p_{2}-1}...x_{n}^{p_{n}-1}\mathrm{d}x_{1}\mathrm{d}x_{2}...\mathrm{d}x_{n}</math>

<math>=\frac{\Gamma\left(p_{1}\right)\Gamma\left(p_{2}\right)...\Gamma\left(p_{n-1}\right)}{\Gamma\left(p_{1}+p_{2}+...+p_{n-1}\right)}\int_{0}^{1}\mathrm{d}s\int_{0}^{1-s}f\left(s+x_{n}\right)s^{p_{1}+p_{2}+...+p_{n-1}-1}x_{n}^{p_{n}-1}\mathrm{d}x_{n}</math>

<math>=\frac{\Gamma\left(p_{1}\right)\Gamma\left(p_{2}\right)...\Gamma\left(p_{n-1}\right)}{\Gamma\left(p_{1}+p_{2}+...+p_{n-1}\right)}\iint_{s,x_{n}\geqslant0;s+x_{n}\leqslant1}f\left(s+x_{n}\right)s^{p_{1}+p_{2}+...+p_{n-1}-1}x_{n}^{p_{n}-1}\mathrm{d}x_{n}</math>

<math>=\frac{\Gamma\left(p_{1}\right)\Gamma\left(p_{2}\right)...\Gamma\left(p_{n-1}\right)}{\Gamma\left(p_{1}+p_{2}+...+p_{n-1}\right)}\cdot\frac{\Gamma\left(p_{1}+p_{2}+...+p_{n-1}\right)\Gamma\left(p_{n}\right)}{\Gamma\left(p_{1}+p_{2}+...+p_{n}\right)}\int_{0}^{1}f\left(u\right)u^{p_{1}+p_{2}+...+p_{n}-1}\mathrm{d}u</math>

<math>=\frac{\Gamma\left(p_{1}\right)\Gamma\left(p_{2}\right)...\Gamma\left(p_{n}\right)}{\Gamma\left(p_{1}+p_{2}+...+p_{n}\right)}\int_{0}^{1}f\left(u\right)u^{p_{1}+p_{2}+...+p_{n}-1}\mathrm{d}u</math>

证明完毕。<ref name="a1"></ref>

==参考资料==
{{Reflist}}

[[Category:积分学]]