I am looking to simplify these term [ I forgot the 3 :( ]


where $a$ and $b$ are two non-negative reals.

(This is not homework. I am just trying to make my expression easy, but I didn't find a way.)

Thanks for your help.

  • $\begingroup$ by simple I mean not a quotient, Thanks for all the answers but I have the feeling that it can be just a constant!!!!! I am not sure $\endgroup$ – Bernstein Jun 28 '19 at 15:36
  • $\begingroup$ I don't think that's possible. You can eliminate either $\sin$ or $\cos,$ but that's about the limit. Perhaps you could provide some larger context for helping us understand what you're trying to do? $\endgroup$ – Adrian Keister Jun 28 '19 at 15:43
  • $\begingroup$ this is the quotient of a radius of an ellipse over the radius his rotation by $\pi/2$ $\endgroup$ – Bernstein Jun 28 '19 at 15:58
  • $\begingroup$ It can't be constant. Geometrically -- for positive $a$ and $b$ -- it's the quotient of the distances from the origin of two ellipses with axis $a$, $b$ and $b$, $a$, resp. in corresponding times $t$. $\endgroup$ – Michael Hoppe Jun 28 '19 at 16:00
  • $\begingroup$ yes you are right $\endgroup$ – Bernstein Jun 28 '19 at 16:01

The form is $$\sqrt{\frac{f^3}{g^3}}$$ which we can write as $$\left(\frac{f}{g}\right)^{3/2}$$ so let's just worry about that inner quotient, $f/g$.

The quotient is definitely not constant; different $t$ values give different results: $$t = 0 \;\to\; \frac{a^2}{b^2} \qquad\qquad t= \frac{\pi}{2}\;\to\;\frac{b^2}{a^2}$$

Having to set aside this dream case, we're left to consider a few alternatives and decide which might be consider least-bad.

  • Leave it as-is. $$\frac{a^2\cos^2 t + b^2 \sin^2 t}{a^2 \sin^2 t+b^2\cos^2 t} \tag{1}$$ That expression isn't terribly complicated.

  • Divide-through by $\cos^2t$ in the numerator and denominator, to get $$\frac{a^2+b^2\tan^2t}{b^2+a^2\tan^2t} \tag{2}$$ This may not be appreciably better, though ... and it introduces unnecessary concern about $t=\pi/2$.

  • @Dr.SonnhardGraubner's answer invokes the double-angle formula to get something I'll write as $$\frac{\left(a^2+b^2\right)+\left(a^2-b^2\right)\cos 2t}{\left(a^2+b^2\right)-\left(a^2-b^2\right)\cos 2t} \tag{3}$$ which is "simpler" in that the degree of the trig functions is lower, at the cose of adding some complexity to the coefficients.

  • @Andrei's suggestion to trade $a$ and $b$ for trig functions is a good one, although I'd choose to swap the sine and cosine assignment to write $a = \sqrt{a^2+b^2} \cos u$ (and $b = \sqrt{a^2+b^2} \sin u$). I'd also use the double-angle formulas to simplify the resulting quotient to $$\frac{1 + \cos 2t \cos 2u}{1 - \cos 2t \cos 2u} \tag{4}$$ (I'd also probably choose to associate $a$ with $\cos u$ (and $b$ with $\sin u$), which changes the above slightly.) Note that $(3)$ seems to be crying-out for us to make such a substitution.

  • Since the quotient appears in the context of ellipses, we could use $a^2-b^2=c^2$ (where $c$ is the center-to-focus distance) and $c = ea$ (where $e$ is the eccentricity) to write $$\frac{ 1 -e^2 \sin^2 t}{1 -e^2 \cos^2 t} \tag{5}$$ I like this one, personally.

  • One can also combine re-writing in terms of $e$ with re-writing in terms of $\cos 2t$ (left as an exercise to the reader), but this just seems to re-complicate things.

One can imagine other variations, too. How useful any one version is depends upon how it's intended to be used.


Divide both the numerator and denominator by $a^2+b^2$, then use $\frac{a^2}{a^2+b^2}=\sin^2 u$. Your expression will become $$\sqrt{\frac{\sin^2 u\cos^2 t+\cos^2u\sin^2t}{\cos^2 u\cos^2 t+\sin^2u\sin^2t}}$$ Now use $\sin^2u+\cos^2 u=1$ $$\sqrt{\frac{\sin^2 u\cos^2 t+\cos^2u\sin^2t}{\cos^2 u\cos^2 t+\sin^2u\sin^2t}}=\sqrt{\frac{(1-\cos^2 u)\cos^2 t+(1-\sin^2u)\sin^2t}{\cos^2 u\cos^2 t+\sin^2u\sin^2t}}\\=\sqrt{\frac{(\cos^2 t+\sin^2 t)-(\cos^2 u\cos^2 t+\sin^2u\sin^2t)}{\cos^2 u\cos^2 t+\sin^2u\sin^2t}}\\=\sqrt{\frac{1}{\cos^2 u\cos^2 t+\sin^2u\sin^2t}-1}$$


For the radicand i have got $$\frac{a^2 (-\cos (2 t))-a^2+b^2 \cos (2 t)-b^2}{a^2 \cos (2 t)-a^2-b^2 \cos (2 t)-b^2}$$


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.