Constant Integrand?

Calculus Level 4

1 3 ( arctan ( x 1 + x 2 ) + arctan ( 1 + x 2 x ) ) d x \large \int_{-1}^3 \left(\arctan\left(\dfrac x{1+x^2}\right) + \arctan\left(\dfrac {1+x^2}x \right)\right) \, dx

If the above integral can be expressed as A π B A\pi^B , where A A and B B are natural numbers . Find A + B A+B


The answer is 2.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

3 solutions

Chew-Seong Cheong
Aug 12, 2016

Let α = arctan ( x 1 + x 2 ) \alpha = \arctan \left(\dfrac x{1+x^2} \right) and β = arctan ( 1 + x 2 x ) \beta = \arctan \left(\dfrac {1+x^2}x \right) tan α = x 1 + x 2 \implies \tan \alpha = \dfrac x{1+x^2} and tan β = 1 + x 2 x \tan \beta = \dfrac {1+x^2}x . tan α = 1 tan β = cot β = tan ( π 2 β ) \implies \tan \alpha = \dfrac 1{\tan \beta} = \cot \beta = \tan \left(\frac \pi 2 - \beta \right) , α = π 2 β \implies \alpha = \frac \pi 2 - \beta α + β = π 2 \implies \alpha + \beta = \frac \pi 2 for α > 0 \alpha > 0 and β > 0 \beta > 0 . We note that when x < 0 x < 0 , α , β < 0 \alpha, \beta < 0 , and α + β = π 2 \alpha + \beta = - \frac \pi 2 . Therefore, we have:

I = 1 3 ( arctan ( x 1 + x 2 ) + arctan ( 1 + x 2 x ) ) d x = 1 3 ( α + β ) d x = 1 0 ( α + β ) d x + 0 3 ( α + β ) d x = 1 0 π 2 d x + 0 3 π 2 d x = π 2 + 3 π 2 = π \begin{aligned} I & = \int_{-1}^3 \left( \arctan \left(\dfrac x{1+x^2} \right) + \arctan \left(\dfrac {1+x^2}x \right) \right) \ dx \\ & = \int_{-1}^3 (\alpha + \beta) \ dx \\ & = \int_{-1}^0 (\alpha + \beta) \ dx + \int_0^3 (\alpha + \beta) \ dx \\ & = \int_{-1}^0 - \frac \pi 2 \ dx + \int_0^3 \frac \pi 2 \ dx \\ & = - \frac \pi 2 + \frac {3 \pi}2 \\ & = \pi \end{aligned}

A + B = 1 + 1 = 2 \implies A+B = 1+1 = \boxed{2}

Moderator note:

Careful there, cot β = tan ( π 2 β ) \cot \beta = \tan \left( \frac\pi2 - \beta \right) is indeed true, but so is cot β = tan ( 5 π 2 β ) \cot \beta = \tan \left( \frac{5\pi}2 - \beta \right) . So, why can't we have α + β = 5 π 2 \alpha + \beta = \frac{5\pi}2 ?

Bonus question : Can you find the fallacy in the working below?

Let y = arctan ( x 1 + x 2 ) + arctan ( 1 + x 2 x ) y = \arctan\left(\frac x{1+x^2}\right) + \arctan\left(\frac {1+x^2}x \right) , then d y d x = 0 \frac{dy} { d x } = 0 . So y y is a constant. Substituting x = 1 x = 1 tells us that y y is a constant function and is equal to π 2 \frac\pi2 . Hence, our integral is equal to 1 3 π 2 d x = 2 π \displaystyle \int_{-1}^3 \frac{\pi}2 \, dx = 2\pi .

Response to the moderator's bonus question: arctan((1+x^2)/x) is undefined at x=0, and is therefore indifferentiable. Since it is not continuous on the interval {x|-1≤x≤3}, then sum of arctan(x/(1+x^2 )) and arctan((1+x^2)/x) is also discontinuous at x=0 (and thus in the range {x|-1≤x≤3}). The derivative of y does not give insight into the behavior of the function around x=0. Additionally, arctan(u)+arctan(1/u) = π/2 for positive u, and -π/2 for negative u.

Michael Boyd - 2 years, 6 months ago
Chung Kevin
Aug 10, 2016

Relevant wiki: Inverse Trigonometric Identities

If x > 0 x> 0 , then we can construct a right-angle triangle with base and height of x x and 1 + x 2 1 + x^2 .

Looking at the diagram, we have the triangles angle sum of a + b + π 2 = π a + b + \dfrac\pi2 = \pi , or equivalently, a + b = π 2 a + b = \dfrac\pi 2 .

Because tan a = 1 1 + x 2 \tan a = \dfrac1{1+x^2} and tan b = 1 + x 2 x \tan b = \dfrac{1+x^2} x , then for x > 0 x> 0 , ( arctan ( x 1 + x 2 ) + arctan ( 1 + x 2 x ) ) = a + b = π 2 \left(\arctan\left(\dfrac x{1+x^2}\right) + \arctan\left(\dfrac {1+x^2}x \right)\right) = a + b = \dfrac\pi2 .

Since the integrand (call it g ( x ) g(x) ) in question is an odd function , then the integral in question can be expressed as

1 3 g ( x ) d x = 1 1 g ( x ) d x = 0 + 1 3 g ( x ) d x = 0 + 1 3 π 2 d x = ( 3 1 ) ( π 2 ) = π . \int_{-1}^3 g(x) \, dx = \underbrace{\int_{-1}^1 g(x) \, dx}_{= \; 0} + \int_1^3 g(x) \, dx = 0 + \int_1^3 \dfrac\pi2 \, dx = (3-1)\left (\dfrac\pi2 \right) = \pi \; .

Hence, A = 1 , B = 1 A + B = 2 A = 1,B=1\Rightarrow A+B = \boxed2 .

Note : A common mistake for students to assume is that ( arctan ( x 1 + x 2 ) + arctan ( 1 + x 2 x ) ) \left(\arctan\left(\dfrac x{1+x^2}\right) + \arctan\left(\dfrac {1+x^2}x \right)\right) is always equal to π 2 \dfrac\pi2 .

Moderator note:

Great setup in explaining why the integrand is, suprisingly, not always a constant!

Kishore S. Shenoy
Aug 11, 2016

Let's analyze the function, y = tan 1 1 x y = \tan^{-1} \dfrac 1x

Substitute tan θ = x tan ( π 2 θ ) = 1 x θ = tan 1 x \color{#3D99F6}{\tan \theta = x\\\tan\left(\dfrac \pi 2 - \theta \right) = \dfrac 1x\\\theta = \tan^{-1}x}

So, y = tan 1 [ tan ( π 2 θ ) ] y = \tan^{-1}\left[\tan\left(\dfrac \pi2 - \theta \right)\right]

Case 1 x 0 x\ge 0

θ ( 0 , π 2 ) ( π 2 θ ) ( 0 , π 2 ) tan 1 [ tan ( π 2 θ ) ] = ( π 2 θ ) \Rightarrow \theta \in\left(0,\dfrac\pi 2\right)\\ \Rightarrow \left(\dfrac \pi 2 - \theta \right) \in \left(0,\dfrac\pi 2\right) \\\Rightarrow \tan^{-1}\left[\tan\left(\dfrac \pi2 - \theta \right) \right]= \left(\dfrac \pi2 - \theta \right)

Case 2 x 0 x \le 0

θ ( π 2 , 0 ) ( π 2 θ ) ( π 2 , π ) \Rightarrow \theta \in\left(-\dfrac\pi 2,0\right) \\\Rightarrow \left(\dfrac \pi2 - \theta \right) \in \left(\dfrac\pi 2,\pi\right)

Remember that for ϕ ( π 2 , π ) \phi \in\left(\dfrac\pi 2,\pi\right) , tan 1 ( tan ϕ ) = ϕ π \tan^{-1}\left(\tan \phi\right) = \phi - \pi

So, tan 1 [ tan ( π 2 θ ) ] = ( π 2 θ ) π = π 2 θ \tan^{-1}\left[\tan\left(\dfrac \pi2 - \theta \right) \right] = \left(\dfrac \pi2 - \theta \right) - \pi = -\dfrac \pi2 - \theta

tan 1 1 x = { π 2 tan 1 x for x > 0 π 2 tan 1 x for x < 0 \large \therefore \tan^{-1}\dfrac1x = \begin{cases}\dfrac \pi2 - \tan^{-1}x &\text{for } x>0\\-\dfrac \pi2 - \tan^{-1}x&\text{for } x<0\end{cases}


So our integral, I = 1 3 tan 1 ( x 1 + x 2 ) + tan 1 ( 1 + x 2 x ) d x = 1 0 π 2 d x + 0 3 π 2 d x = π I = \int_{-1}^3 \tan^{-1}\left(\dfrac x{1+x^2}\right)+\tan^{-1}\left(\dfrac {1+x^2}x \right) \, dx = \int_{-1}^0 -\dfrac \pi 2 \,dx + \int_0^3 \dfrac\pi 2 \, dx = \boxed{\pi}

Moderator note:

Yes. This essentially boils down to recognizing that the integrand is a piecewise function.

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...