In the equation $$x\cos(\theta) + y\sin(\theta) = z,$$ how do I solve in terms of $\theta$? i.e $\theta = \dots$.
-
3Linear equations in $\sin \theta$ and $\cos \theta$ can be solved by a resolvent quadratic equation. One method is to write the $\sin \theta$ and $\cos \theta$ functions in terms of the $\tan (\theta/2)$. – Américo Tavares Apr 15 '11 at 15:44
3 Answers
There are various possible strategies. I will mention one approach. Of course if there is a solution, there are infinitely many, since we can add $2\pi$ to any solution and get another solution.
Let's change notation a little. We are interested in the equation $$a\cos\theta + b\sin\theta=q$$
Rewrite this equation as $$\frac{a}{\sqrt{a^2+b^2}}\cos\theta+ \frac{b}{\sqrt{a^2+b^2}}\sin\theta=\frac{q}{\sqrt{a^2+b^2}}$$ Now let $\phi$ be the angle whose sine is $a/\sqrt{a^2+b^2}$ and whose cosine is $b/\sqrt{a^2+b^2}$. By a formula that I hope is familiar (sine of a sum of angles), the equation can be rewritten as $$\sin(\phi+\theta)=\frac{q}{\sqrt{a^2+b^2}}$$ Look at the right-hand side. If its absolute value is greater than $1$, there will be no (real) solution. Otherwise, for simplicity, call the right-hand side $w$. Then we can write that $\phi+\theta=\arcsin w$ or $\phi+\theta=\pi-\arcsin w$. Now remember that whatever solutions you get through this process, anything obtained by adding $2n\pi$, where $n$ is an integer, to a solution, is also a solution.
The reason I went in detail through this approach is that in Physics, it is often important to express $a\cos\theta+b\sin\theta$ in the form $c\sin(\phi+\theta)$ that we used to solve the equation. There are other approaches.

- 507,029
-
1Almost right, but you slipped at the end: $\sin(\phi+\theta)=\sin(w)$ means that $\phi+\theta$ belongs to $\arcsin(w)+2\pi\mathbb{Z}$ or to $\pi-\arcsin(w)+2\pi\mathbb{Z}$. – Did Apr 15 '11 at 16:14
-
You did not get what I wrote: unless $|w|=1$, the set of solutions of the equation $\sin(\phi+\theta)=w$ cannot be written as the set of $\theta_0+2n\pi$, for a well chosen $\theta_0$ and for every integer $n$. So your solution as written is false. – Did Apr 15 '11 at 16:36
-
@Didier Pau: Sorry, I missed part of the content of your comment. I have edited the answer to take account of what you wrote. – André Nicolas Apr 15 '11 at 16:37
This expands my comment above. As I wrote here "certain trigonometric equations such as the linear equations in $\sin x$ and $\cos x$ can be solved by a resolvent quadratic equation. One method is to write the $\sin x$ and $\cos x$ functions in terms of (...) $\tan$ of the half-angle".
Applying this method, since
$$\cos \theta =\frac{1-\tan ^{2}\frac{\theta }{2}}{1+\tan ^{2}\frac{\theta }{2% }}\qquad\text{and}\qquad\sin \theta =\frac{2\tan \frac{\theta }{2}}{1+\tan ^{2}\frac{\theta }{2}},$$
your equation is equivalent to
$$x\frac{1-\tan ^{2}\frac{\theta }{2}}{1+\tan ^{2}\frac{\theta }{2}}+y\frac{% 2\tan \frac{\theta }{2}}{1+\tan ^{2}\frac{\theta }{2}}=z.$$
Let $u=\tan \frac{\theta }{2}$. Then we can write it as
$$\left( x+z\right) u^{2}-2yu+z-x=0,$$
which has the solutions
$$u=\tan \frac{\theta }{2}=\frac{1}{2\left( x+z\right) }\left( 2y\pm2\sqrt{% y^{2}+x^{2}-z^{2}}\right).$$
Thus
$$\theta =2\arctan \left( \frac{1}{ x+z }\left( y\pm % \sqrt{y^{2}+x^{2}-z^{2}}\right) \right) +2n\pi,\qquad n\in\mathbb{Z} .$$
This method is valid iff $\theta \neq (2k+1)\pi $, with $k\in\mathbb{Z}$.
A different technique is to use an auxiliary angle.

- 38,602
-
-
You are welcome! I adapted this technique from my 1967 Trigonometry text book (J. Calado, Compêndio de Trigonometria), where it is presented and worked out in detail. I deleted my explanation of the auxiliary angle technique (answered before by user6312). – Américo Tavares Apr 15 '11 at 21:13
Let us introduce $c=\cos \theta$. Then your equation reads $$x c + s y \sqrt{1-c^2} =z$$ where $s=\pm 1$ is related to the quadrant of $\theta$ (+1 in the first and second quadrant and -1 in the third and forth). Subtracting $x c$ from both sides then squaring the equation yields $$y^2 (1-c^2) = (z- xc)^2 = z^2 -2 zx c + x^2 c^2.$$ This is a quadratic equation with the solutions $$ c_\pm = \frac{x z \pm y \sqrt{x^2+ y^2 - z^2}}{x^2 + y^2}.$$ In order that $c_\pm$ are real, we need $z^2 \leq x^2+y^2$. As we have squared the equation, we have to check whether $c_\pm$ solves the original equation. Indeed, $c_\pm$ solves the original equation with $s_\pm=\text{sgn}[(z-xc_\pm)/y]$. Therefore, we have the two solutions (mod $2\pi$) $$\theta = s_\pm \arccos(c_\pm). $$

- 23,360
-
Hmmm... I doubt that. First, both solutions $c_+$ and $c_-$ of the quadratic equation are admissible. Second, the set of solutions $\theta$ is either empty or of the form $a\pm b+2\pi\mathbb{Z}$. – Did Apr 15 '11 at 16:10
-
@Didier Piau: thank you, I corrected the mistake. Regarding the $+2\pi \mathbb{Z}$, as $\theta$ is an angle everything is mod $2\pi$. – Fabian Apr 15 '11 at 16:18
-
Did you check that $|c_\pm|\le1$? You know, me saying so should not be enough... And I am not sure I follow your dealing with the signs. – Did Apr 15 '11 at 16:31
-
@Didier Piau: yes I check it (I actually started writing the edited post before I read your comment...) – Fabian Apr 15 '11 at 16:33
-
@Didier Piau: the signs are easy. When I square the equation, I have a √ on one side and $(z-xc)/sy$ on the other. The latter should be positive for the equality to hold. – Fabian Apr 15 '11 at 16:36
-
The signs one has to avoid to get values which can be solutions are easy to determine (this is your last comment). But the problem is to be sure that the remaining ones actually yield solutions (this was my remark). (In case you are wondering, I know how to solve this exercise, thank you. My point is, too many steps in your post are left vague for it to be convincing.) – Did Apr 15 '11 at 16:47