Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I am facing a problem and I do not know if it is solvable or not.

Suppose I have 2 points and a distance, $P_1$, $P_2$ and $D_x$ respectively.

I need a mathematical way to find the center of a circle $\mathcal{C}_1$ that satisfies the following:

  1. The center of $\mathcal{C}_1$ is on the y-axis
  2. $P_1$ is on the circumference $\mathcal{C}_1$
  3. $\mathcal{C}_1$ is tangent to circle $\mathcal{C}_2$, where $\mathcal{C}_2$'s center is $P_2$ and its radius is $D_x$.

I tried a lot to find it but I could not.

share|improve this question
add comment

1 Answer

up vote 0 down vote accepted

If $\mathcal C_1$ with radius $r$ is tangent to $\mathcal C_2$, then $\mathcal C_1$ is at distance $|r\pm D_x|$ from $P_2$. Thus we can rephrase the problem as finding a point on the $y$ axis at distance $r$ from $P_1$ and $|r\pm D_x|$ from $P_2$. Let $(x_1,y_1)$, $(x_2,y_2)$ and $(0,y)$ denote the coordinates of $P_1$, $P_2$ and the centre of $\mathcal C_1$, respectively. Then

$$ \begin{align} x_1^2+(y-y_1)^2&=r^2\;,\\ x_2^2+(y-y_2)^2&=(r\pm D_x)^2\;. \end{align} $$

Subtracting the two equations yields a linear relationship between $y$ and $r$:

$$ x_2^2+y_2^2-(x_1^2+y_1^2)+2y(y_1-y_2)=D_x^2\pm2D_xr\;. $$

For given $y$, exactly one of the solutions for $r$ is non-negative. Solving for $r$ and substituting into the first equation yields

$$ 4D_x^2\left(x_1^2+(y-y_1)^2\right)=\left(x_2^2+y_2^2-(x_1^2+y_1^2)+2y(y_1-y_2)-D_x^2\right)^2\;. $$

According to Wolfram|Alpha, the discriminant of this quadratic equation in $y$ has three roots with respect to $D_x^2$. The one at $D_x^2=0$ is clear: If the circle degenerates to a point, the two solutions coalesce. The other two roots have a nice interpretation. Here $D_x$ is the distance from $P_2$ to $P_1$ and to the reflection of $P_1$ in the $y$ axis, respectively. Any circle with centre on the $y$ axis that goes through $P_1$ also goes through the reflection of $P_1$ in the $y$ axis. If one of these two points is inside $\mathcal C_2$ and the other is outside, a circle through both of them necessarily intersects $\mathcal C_2$, so in this case the problem has no solution. There are two solutions if $P_1$ and its reflection in the $y$ axis either both lie inside or both lie outside $\mathcal C_2$. If one of them lies exactly on $\mathcal C_2$, there is one solution, with the point of tangency coinciding with the point that lies on $\mathcal C_2$, and the centre of $\mathcal C_1$ given by the intersection of the $y$ axis with the line containing the radius of $\mathcal C_2$ through that point.

share|improve this answer
    
Thank you! You made my day. –  mohamed ehab Sep 20 '12 at 10:22
    
@mohamed: You're welcome! –  joriki Sep 20 '12 at 10:24
    
but I do not understand something, In wolfram's link, $\mathcal k$ is considered the variable, not $\mathcal y$.... right ? If this is true, I already know $\mathcal k$ (which is $\mathcal D_x$ in my case), but I do not know $\mathcal y$. –  mohamed ehab Sep 20 '12 at 10:48
    
@mohamed: I think that's a misunderstanding. The discriminant is the discriminant with respect to $y$ (indicated by the "$\{y\}$" at the end of the "Input" field), and then the roots of that discriminant with respect to $k$ are given. These are the values of $k$ at which the quadratic equation in $y$ switches from two solutions through one solution to zero solutions. That is, if you fix $P_1$ and $P_2$ and let $k$ vary, there are two solutions for $y$ either if $k$ is small enough that both $P_1$ and its reflection are outside $\mathcal C_2$, or if $k$ is large enough that they're both inside. –  joriki Sep 20 '12 at 11:04
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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