From conjugate semi-diameters to semi-axes

Geometry Level 5

Consider the ellipse described by

r ( t ) = C + f 1 cos t + f 2 sin t \mathbf{r}(t) = \mathbf{C} + \mathbf{f_1} \cos t + \mathbf{f_2} \sin t

The vectors f 1 , f 2 \mathbf{f_1} , \mathbf{f_2} are a pair of conjugate semi-diameters of this ellipse.

Now, suppose you are given that f 1 = ( 3 , 1 ) \mathbf{f_1} = (3, 1) and f 2 = ( 0 , 2 ) \mathbf{f_2} = (0, 2) . Find the sum of the semi-axes lengths of this ellipse.


The answer is 5.099.

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

Abhishek Sinha
Jul 11, 2020

Translating the center to the point C C , we can write down the equation of the ellipse in the parametric form as: x ( t ) = 3 cos ( t ) , y ( t ) = cos ( t ) + 2 sin ( t ) . x(t) = 3 \cos(t), y(t) = \cos(t) + 2 \sin(t). Hence, in polar coordinates, the length of the radius vector is given by r ( t ) = x 2 ( t ) + y ( t ) 2 = 7 + 13 cos ( 2 t + α ) , r(t) = \sqrt{x^2(t)+ y(t)^2}= \sqrt{7+ \sqrt{13}\cos(2t+ \alpha)}, for some angle α \alpha . The length of the semi-major and semi-monor axes are given by the maximum and minimum values of r ( t ) r(t) respectively. Hence, their sum is 7 + 13 + 7 13 5.099. \sqrt{7+ \sqrt{13}} + \sqrt{7- \sqrt{13}} \approx 5.099.

Mark Hennings
Jul 11, 2020

We can ignore c \mathbf{c} , since it just shifts the centre of the ellipse. Thus we have x = 3 cos t y = cos t + 2 sin t x \; =\; 3\cos t \hspace{2cm} y \; = \; \cos t + 2\sin t and hence 36 = 4 x 2 + ( 3 y x ) 2 = 5 x 2 6 x y + 9 y 2 = ( x y ) ( 5 3 3 9 ) ( x y ) = ( x y ) P T ( u 0 0 v ) P ( x y ) 36 \; = \; 4x^2 + (3y - x)^2 \; = \; 5x^2 - 6xy + 9y^2 \; = \; \big(\,x\;\;y\,\big)\left(\begin{array}{cc} 5 & -3 \\ -3 & 9 \end{array}\right)\left(\begin{array}{c} x \\ y \end{array}\right) \; = \; \big(\,x\;\;y\,\big)P^T\left(\begin{array}{cc} u & 0 \\ 0 & v \end{array}\right)P\left(\begin{array}{c} x \\ y \end{array}\right) for some orthogonal matrix P P , where u , v u,v are the eigenvalues of the matrix ( 5 3 3 9 ) \left(\begin{array}{cc} 5 & -3 \\ -3 & 9 \end{array}\right) Thus the semimajor and semiminor axes a , b a,b are 6 u \tfrac{6}{\sqrt{u}} and 6 v \tfrac{6}{\sqrt{v}} , and hence we want to find a + b = 6 u + 6 v = 6 ( u + v ) u v a + b \; =\; \frac{6}{\sqrt{u}} + \frac{6}{\sqrt{v}} \; = \; \frac{6(\sqrt{u}+\sqrt{v})}{\sqrt{uv}} noting that ( a + b ) 2 = 36 u v ( u + v + 2 u v ) (a+b)^2 \; = \; \frac{36}{uv}\big(u + v + 2\sqrt{uv}\big) Now u , v u,v are the roots of the characteristic polynomial ( t 5 ) ( t 9 ) 9 = t 2 14 t + 36 (t-5)(t-9)-9 \; = \; t^2 - 14t + 36 so that u + v = 14 , u v = 36 u+v=14,uv=36 and hence ( a + b ) 2 = 14 + 2 × 6 = 26 (a+b)^2 = 14 +2\times6 =26 making the answer 26 \boxed{\sqrt{26}} .

David Vreken
Jul 10, 2020

For an ellipse p ( t ) = f 0 + f 1 cos t + f 2 sin t \overrightarrow{p}(t) = \overrightarrow{f_0} + \overrightarrow{f_1} \cos t + \overrightarrow{f_2} \sin t , the four vertices are p ( t 0 ) \overrightarrow{p}(t_0) , p ( t 0 ± π 2 ) \overrightarrow{p}(t_0 \pm \frac{\pi}{2}) , and p ( t 0 + π ) \overrightarrow{p}(t_0 + \pi) where t 0 t_0 is defined by cot ( 2 t 0 ) = f 1 2 f 2 2 2 f 1 f 2 \cot (2t_0) = \frac{\overrightarrow{f_1}^2 - \overrightarrow{f_2}^2}{2\overrightarrow{f_1} \cdot \overrightarrow{f_2}} (see here ).

Using f 1 = ( 3 , 1 ) \overrightarrow{f_1} = (3, 1) and f 2 = ( 0 , 2 ) \overrightarrow{f_2} = (0, 2) , cot ( 2 t 0 ) = ( 3 2 + 1 2 ) ( 2 2 0 2 ) 2 ( 3 0 + 1 2 ) = 3 2 \cot (2t_0) = \frac{(3^2 + 1^2) - (2^2 - 0^2)}{2(3 \cdot 0 + 1 \cdot 2)} = \frac{3}{2} or tan ( 2 t 0 ) = 2 3 \tan (2t_0) = \frac{2}{3} , and using some trig identities:

tan t 0 = 3 + 13 2 \tan t_0 = \frac{-3 + \sqrt{13}}{2}

sin t 0 = 13 3 13 26 \sin t_0 = \sqrt{\frac{13-3\sqrt{13}}{26}}

cos t 0 = 13 + 3 13 26 \cos t_0 = \sqrt{\frac{13+3\sqrt{13}}{26}}

Since the endpoints of the major axis are defined by p ( t 0 ) \overrightarrow{p}(t_0) and p ( t 0 + π ) \overrightarrow{p}(t_0 + \pi) :

semi-major axis a a

= 1 2 p ( t 0 ) p ( t 0 + π ) = \frac{1}{2}|\overrightarrow{p}(t_0) - \overrightarrow{p}(t_0 + \pi)|

= 1 2 ( f 0 + ( 3 , 1 ) cos t 0 + ( 0 , 2 ) sin t 0 ) ( f 0 + ( 3 , 1 ) cos ( t 0 + π ) + ( 0 , 2 ) sin ( t 0 + π ) ) = \frac{1}{2}|(\overrightarrow{f_0} + (3, 1) \cos t_0 + (0, 2) \sin t_0) - (\overrightarrow{f_0} + (3, 1) \cos (t_0 + \pi) + (0, 2) \sin (t_0 + \pi))|

= ( 3 , 1 ) cos t 0 + ( 0 , 2 ) sin t 0 = |(3, 1) \cos t_0 + (0, 2) \sin t_0|

= ( 3 , 1 ) 13 + 3 13 26 + ( 0 , 2 ) 13 3 13 26 = |(3, 1) \sqrt{\frac{13+3\sqrt{13}}{26}} + (0, 2) \sqrt{\frac{13-3\sqrt{13}}{26}}|

= ( 3 13 + 3 13 26 ) 2 + ( 1 13 + 3 13 26 + 2 13 3 13 26 ) 2 = \sqrt{(3 \cdot \sqrt{\frac{13+3\sqrt{13}}{26}})^2 + (1 \cdot \sqrt{\frac{13+3\sqrt{13}}{26}} + 2 \cdot \sqrt{\frac{13-3\sqrt{13}}{26}})^2}

= 7 + 13 = \sqrt{7 + \sqrt{13}}

Since the endpoints of the minor axis are defined by p ( t 0 ± π 2 ) \overrightarrow{p}(t_0 \pm \frac{\pi}{2}) :

semi-minor axis b b

= 1 2 p ( t 0 π 2 ) p ( t 0 + π 2 ) = \frac{1}{2}|\overrightarrow{p}(t_0 - \frac{\pi}{2}) - \overrightarrow{p}(t_0 + \frac{\pi}{2})|

= 1 2 ( f 0 + ( 3 , 1 ) cos ( t 0 π 2 ) + ( 0 , 2 ) sin ( t 0 π 2 ) ) ( f 0 + ( 3 , 1 ) cos ( t 0 + π 2 ) + ( 0 , 2 ) sin ( t 0 + π 2 ) ) = \frac{1}{2}|(\overrightarrow{f_0} + (3, 1) \cos (t_0 - \frac{\pi}{2}) + (0, 2) \sin (t_0 - \frac{\pi}{2})) - (\overrightarrow{f_0} + (3, 1) \cos (t_0 + \frac{\pi}{2}) + (0, 2) \sin (t_0 + \frac{\pi}{2}))|

= ( 3 , 1 ) sin t 0 ( 0 , 2 ) cos t 0 = |(3, 1) \sin t_0 - (0, 2) \cos t_0|

= ( 3 , 1 ) 13 3 13 26 + ( 0 , 2 ) 13 + 3 13 26 = |(3, 1) \sqrt{\frac{13-3\sqrt{13}}{26}} + (0, 2) \sqrt{\frac{13+3\sqrt{13}}{26}}|

= ( 3 13 3 13 26 ) 2 + ( 1 13 3 13 26 + 2 13 + 3 13 26 ) 2 = \sqrt{(3 \cdot \sqrt{\frac{13-3\sqrt{13}}{26}})^2 + (1 \cdot \sqrt{\frac{13-3\sqrt{13}}{26}} + 2 \cdot \sqrt{\frac{13+3\sqrt{13}}{26}})^2}

= 7 13 = \sqrt{7 - \sqrt{13}}

Therefore, the sum of the semi-axes lengths is a + b = 7 + 13 + 7 13 5.099 a + b = \sqrt{7 + \sqrt{13}} + \sqrt{7 - \sqrt{13}} \approx \boxed{5.099} .


Using the same method above with f 1 = ( f 1 x , f 1 y ) \overrightarrow{f_1} = (f_{1x}, f_{1y}) and f 2 = ( f 2 x , f 2 y ) \overrightarrow{f_2} = (f_{2x}, f_{2y}) , the semi-axes lengths are:

1 2 ( f 1 x 2 + f 1 y 2 + f 2 x 2 + f 2 y 2 ) ± 1 2 ( ( f 1 x f 2 y ) 2 + ( f 1 y + f 2 x ) 2 ) ( ( f 1 x + f 2 y ) 2 + ( f 1 y f 2 x ) 2 ) \sqrt{\frac{1}{2}(f_{1x}^2 + f_{1y}^2 + f_{2x}^2 + f_{2y}^2) \pm \frac{1}{2}\sqrt{((f_{1x} - f_{2y})^2 + (f_{1y} + f_{2x})^2)((f_{1x} + f_{2y})^2 + (f_{1y} - f_{2x})^2)}}

and the sum of the semi-axes lengths are:

max ( ( f 1 x + f 2 y ) 2 + ( f 1 y f 2 x ) 2 , ( f 1 x f 2 y ) 2 + ( f 1 y + f 2 x ) 2 ) \max(\sqrt{(f_{1x} + f_{2y})^2 + (f_{1y} - f_{2x})^2}, \sqrt{(f_{1x} - f_{2y})^2 + (f_{1y} + f_{2x})^2})

Of course 14 ± 2 13 = 13 ± 1 \sqrt{14 \pm 2\sqrt{13}} = \sqrt{13} \pm 1 , and hence 7 + 13 + 7 13 = 13 + 1 2 + 13 1 2 = 26 \sqrt{7 + \sqrt{13}} + \sqrt{7 - \sqrt{13}} \; = \; \frac{\sqrt{13}+1}{\sqrt{2}} + \frac{\sqrt{13}-1}{\sqrt{2}} \; = \; \sqrt{26} You have a small typo in the very last line of your proof (one + + should be - ).

Mark Hennings - 11 months ago

Log in to reply

Thanks, I edited it! I also didn't notice that 7 + 13 + 7 13 = 26 \sqrt{7 + \sqrt{13}} + \sqrt{7 - \sqrt{13}} = \sqrt{26} . Nice observation!

David Vreken - 11 months ago

Using the same method above with f 1 = ( f 1 x , f 1 y ) \overrightarrow{f_1} = (f_{1x}, f_{1y}) and f 2 = ( f 2 x , f 2 y ) \overrightarrow{f_2} = (f_{2x}, f_{2y}) , the semi-axes lengths are 1 2 ( f 1 x 2 + f 1 y 2 + f 2 x 2 + f 2 y 2 ) ± 1 2 ( ( f 1 x f 2 y ) 2 + ( f 1 y + f 2 x ) 2 ) ( ( f 1 x + f 2 y ) 2 + ( f 1 y f 2 x ) 2 ) \sqrt{\frac{1}{2}(f_{1x}^2 + f_{1y}^2 + f_{2x}^2 + f_{2y}^2) \pm \frac{1}{2}\sqrt{((f_{1x} - f_{2y})^2 + (f_{1y} + f_{2x})^2)((f_{1x} + f_{2y})^2 + (f_{1y} - f_{2x})^2)}} .

David Vreken - 11 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...