Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

For what real values of $\alpha$ does this integral converge?

$$\int_0^\infty {x^\alpha e^{-x} dx}$$

And what's the usual approach to this kind of problem?

share|improve this question
2  
en.wikipedia.org/wiki/Gamma_function –  Jack Nov 25 '12 at 21:18
    
@Jack: But my textbook says it's only convergent when $\alpha>-1$, which differs with what Wikipedia says ($\alpha\not\in\mathbb{Z}_{(-\infty,-1]}$). –  Voldemort Nov 25 '12 at 21:27
add comment

3 Answers

up vote 3 down vote accepted

Note: $$\frac{1}{e} \int_0^1 {x^\alpha dx}<\int_0^1 {x^\alpha e^{-x} dx} <\int_0^1{x^\alpha dx} $$ Thus, $\int_0^1 {x^\alpha e^{-x} dx}$ converges iff $\int_0^1{x^\alpha dx}$ does which happens for $\alpha>-1$. $ \int_1^\infty {x^\alpha e^{-x}dx} < \int_1^\infty {x^\alpha ( x^N/N!)^{-1} dx}$ (where $N>\alpha +2$) does always converge.

share|improve this answer
    
+1. Neat. ${}{}$ –  user17762 Nov 25 '12 at 21:34
    
But why in the Gamma Function article wikipedia says that it can also be defined at all points except $\mathbb{Z}_{(-\infty,-1]}$? –  Voldemort Nov 25 '12 at 21:37
1  
Wikipedia says that you can continue the function defined by this to an analytical function defined everywhere except on the negative integers but not that the integral itself converges everywhere. –  Dominik Nov 25 '12 at 21:41
add comment

Note that for $\alpha > -1$, there exists $X(\alpha)$ such that $x^{\alpha} e^{-x} \leq e^{-x/2}$ for all $x > X(\alpha)$.

Hence, $$\int_0^{\infty} x^{\alpha} e^{-x} dx \leq \int_0^{X(\alpha)} x^{\alpha} e^{-x} dx + \int_{X(\alpha)}^{\infty} e^{-x/2} dx = \underbrace{\int_0^{X(\alpha)} x^{\alpha} e^{-x} dx}_{< \infty} + 2e^{-X(\alpha)/2}$$

For $\alpha \leq -1$, the integral diverges, which can be shown as follows. Let $\alpha = -1-\beta$ $$\int_0^{\infty} x^{\alpha} e^{-x} dx = \underbrace{\int_0^{\epsilon} \dfrac{e^{-x}}{x^{1+\beta}} dx}_{I} + \underbrace{\int_{\epsilon}^{\infty} \dfrac{e^{-x}}{x^{1+\beta}} dx}_{II}$$ Note that $$II \leq \displaystyle \int_{\epsilon}^{\infty} \dfrac1{x^{1+\beta}} dx < \infty$$ Note that $$I \geq \int_0^{\epsilon} \dfrac{e^{-\epsilon}}{x^{1+\beta}} dx = \infty$$ Hence, the integral converges for $\alpha > 1$ and diverges for $\alpha \leq -1$.

The integral you are interested in, wherever it converges, is called the Gamma function. The Gamma function is defined for $\alpha \leq -1$ as well through analytic continuation of the above integral. However, the Gamma function still has poles (blows to $\infty$) at negative integers.

share|improve this answer
add comment

I would split the interval into $[0,1] \cup (1,\infty)$.

In $(1,\infty)$, $e^x$ dominates $x^{\alpha}$ as $x \to \infty$, so $\int_1^\infty {x^\alpha e^{-x} dx}$ is convergent for all $\alpha$. (To make this precise: given $\alpha \in \mathbb R$, we can find $x_0$ such that $x^{\alpha} < e^{x/2}$ for all $x \ge x_0$. So $0 \le x^{\alpha}e^{-x} \le e^{-x/2}$ for all $x \ge x_0$, and the integral converges because the integral of $e^{-x/2}$ converges.)

And in $[0,1]$, $0.3 < e^{-x} \le 1$, so the convergence of the integral is determined only by the convergence of $\int_0^1 {x^\alpha dx}$, which you can work out for yourself.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.