Please consider the following case:

$$I = \int^1_{-1}x^2dx$$ $$u(x) = x^2 \rightarrow du = 2x\,dx$$ $$u(-1) = 1, u(1) = 1$$


$$I = \int^1_1\frac{u}{2\sqrt u} du = 0$$

Obviously the problem here is to only consider the positive root of u. I don't know how to consider both roots. This example is trivial but I have another example where such substitution would be really helpful:

$$I = \int^1_{-1}\frac{x^2(1 - x^2)^\frac{3}{2}}{3} - \frac{x^2(1 - x^2)^\frac{5}{2}}{5} - \frac{x^4(1 - x^2)^\frac{3}{2}}{3} dx$$

I don't want you to solve it for me using another method, I know how to use an integral solver online. My question is how to properly do the change of variable.


In general you need to use a one-to-one change of variable; thus for instance if you were to compute $\int_{-1}^1 x^2 dx$ you would split into $\int_{-1}^0 x^2 dx + \int_0^1 x^2 dx$. Then each of those integrals, after substitution and simplification, becomes $\int_0^1 \frac{1}{2} u^{1/2} du$.

(That said, there are some situations where by basically breaking the rules twice you cancel out your mistakes and get the right answer; see Why can the trig sub $x=\cosh(\theta)$ be used to solve the integral: $\int \frac{1}{(x^2-1)^{3/2}}dx$? for instance.)

  • $\begingroup$ It is not necessary for changes of variable to be one-to-one, see my comment here: math.stackexchange.com/a/1101421/5316 $\endgroup$ – Matthew Towers Dec 9 '18 at 19:39
  • 1
    $\begingroup$ @MatthewTowers Yes, it is true that $\int_a^b f'(x) dx$ is always $f(b) - f(a)$, regardless of injectivity, which means you can do various similar things also regardless of injectivity. What injectivity buys you is a lack of a singularity in $\frac{dx}{du}$, which is a serious problem for, say, $u=x^2$ on $[a,b]$ with $a<0,b>0$. $\endgroup$ – Ian Dec 9 '18 at 19:52

While it's certainly true that $du = 2x\,dx$ over the whole region of integration, it's not true that $x = \sqrt{u}$ over the region. This is only true for $x > 0$, so you can't apply it to the whole integral. Instead, you need to split it apart and use the correct transformation $x = -\sqrt{u}$ for the region $x < 0$.

In general, if your change of variables $u(x)$ has a multivalued inverse, you'll need to split the integral into the regions where each possible inverse function $x(u)$ applies.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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