There shouldn't be brackets of any sorts at all actually.Inquiry: Shouldn't the a and b be in braces instead of parentheses? Or does it actually not matter.
There shouldn't be brackets of any sorts at all actually.
Also, the integral diverges.
Nice and easy one for the 2016ers
Whoops, I made a silly mistake! Cheers for pointing that out
tan(A-B)=(tan(A)-tan(B))/(1+tan(A)tan(B))
So if A-B is in (-pi/2,pi/2), we have:
A-B=arctan((tan(A)-tan(B))/(1+tan(A)tan(B)))
Let A=arctan(a) and B=arctan(b) and this simplifies to
arctan(a)-arctan(b)=arctan((a-b)/(1+ab))=arccot((1+ab)/(a-b))
Note again that this ia only valid when |LHS| < pi/2.
You can find a general expression casewise for arccot((1+ab)/(a-b)) when this criteria is not satisfied if you like, but I am too lazy to do this right now and we don't need this for b).
b) Let a=x+1, b=x and use our identity, noting that 0 < arctan(x+1)-arctan(x) < 2*arctan(1/2) < pi/2 for all x.
This reduces our integrand to arctan(x+1)-arctan(x).
So the end answer is 1/(1+(x+1)^2)-1/(1+x^2)+const.
You can combine the fractions if you want, but the resulting expresion is not much nicer.
Oh lol, yeah. Got a bit complacent there .
Other method is to just rationalise the numerator straight away and then have a square root of quadratic in the denominator and a linear term in the numerator, which can easily be split up and integrated (bit tedious).Alternatively
Yes, I think making the substitution makes it easier to see.Other method is to just rationalise the numerator straight away and then have a square root of quadratic in the denominator and a linear term in the numerator, which can easily be split up and integrated (bit tedious).
It can't be that messy... can it? My intuition tells me there is a very simple way to do it, and it is killing me because I don't see anything that can make it a three-liner.Im guessing