2

This is from a math competition so it must not be something really long If a parabola touches the lines $y=x$ and $y=-x$ at $A(3,3) $ and $b(1,-1)$ respectively, then

(A) equation of axis of parabola is $2x+y=0$

(B)slope of tangent at vertex is $1/2$

(C) Focus is $(6/5,-3/5)$

(D) Directrix passes through $(1,-2)$

I thought the axis would be the angle bisector of the tangents passing through the focus but it turns out that is not the case in a parabola so how can I find anything..

Jean Marie
  • 81,803
  • Does ''touches'' means that the parabola is tangent to the lines? – Emilio Novati Oct 20 '16 at 15:09
  • The task has 4 conditions: 2 points and two slopes. The "standard" parabola has only 3 free coefficients. I checked that the 4 conditions CANNOT be met. Hence, we have here a tilted parabola which has a 4th free parameter, the tilting angle. Then it works. For tilted parabola see http://math.stackexchange.com/questions/1045286/ – Andreas Oct 20 '16 at 15:46
  • @bulbasaur Problem in which competition? – Narasimham Oct 21 '16 at 14:56
  • @Andreas I agree with the count of degrees of freedom. But the "tilt angle" is usually not a good parameter to think of and to work with for practical determination of parabolas in particular. – Jean Marie Oct 22 '16 at 12:43
  • @JeanMarie This comment was written early - at that point in time 3 (wrong) answers were present only for the non-tilted parabola. I was pointing out the tilt to make people aware of this extra feature which must have been overlooked then. The tilt, by the way, could be used here directly: employ an unknown tilt angle and tilt the lines $y = \pm x$ (very easy) and the two points (still not too difficult) and then solve for a standard parabola. I didn't write down that path to a solution. – Andreas Oct 23 '16 at 16:07
  • I understand and I agree. – Jean Marie Oct 23 '16 at 16:12
  • @Narasimham math quiz from 11th grade.. – bulbasaur Oct 23 '16 at 16:13

4 Answers4

2

This figure illustrate the situation.

enter image description here

Using the fact that:

Tangents drawn at the endpoints of a focal chord of a parabola intersect at right angles (on the directrix).

we can say that the focal chord has equation: $y=2x-3$, so we can test that the answer (C): $F=(6/5,-3/5)$ is correct. So we have the focus.

We can also use the fact that:

The tangent at any point of the parabola is equally inclined to the focal distance and the axis of the parabola.

In the figure this mens that the two angles in $A$ and $D$ are equals and from this we can find the axis.

Finally, from the focus and the axis we can find the directrix as the line orthogonal to axis that passes thorough the common point of the two tangents : $C$.

If you o this you can verify that also the answer (D) is correct.

You can find the properties of the tangent used in this answer at: http://www.nabla.hr/CS-ParabolaAndLine2.htm.

Emilio Novati
  • 62,675
  • Very nice figure [+1]. I am going to give an answer based on an analytical treatment of the relationship focus-directrix : $MF^2=MH^2$. – Jean Marie Oct 20 '16 at 22:02
  • Nice answer but how can you say for sure that the point $(6/5,-3/5)$ is the focus not just a random point lying on the line $y=2x-3$ – bulbasaur Oct 21 '16 at 04:18
  • @Emilio Novati Did you satisfy conditions (A) and (B)? – Narasimham Oct 21 '16 at 13:15
  • @bulbasaur: good point! I guess I hoped that there was not a similar trick in the responses. And I was convinced using this point and seeing that also the answer (D) was correct. But clearly this is not a proof...... – Emilio Novati Oct 21 '16 at 14:36
  • @Narasimham. It seems that (A) and (B) are not compatible with (D). – Emilio Novati Oct 21 '16 at 14:37
  • Yes, (A) and (B) are not compatible with second line given slope and tangent point, and also (C) and (D). So clearly wrong. Your diagram also indicates the same. – Narasimham Oct 21 '16 at 14:51
2

(see figure below)

Parametric equations for the parabola are readily obtained, assuming a certain knowledge of quadratic Bezier curves : see for example the paragraph "Second order curve is a parabolic segment" in (https://en.wikipedia.org/wiki/B%C3%A9zier_curve)).

The tangents' intersection is clearly $O(0,0)$. Thus the parabola is nothing but the Bezier curve with endpoints $A$ and $B$ and "directing" point in $O$: $M=A(1-t)^2 + 2t(1-t)O + B t^2$, otherwise said:

$$\tag{1}\pmatrix{x\\y}=(1-t)^2\pmatrix{3\\3}+2t(1-t)\pmatrix{0\\0}+t^2\pmatrix{1\\-1}$$

$$\tag{2}\Leftrightarrow \ \ \ \ \cases{x=3(1-t)^2+t^2\\y=3(1-t)^2-t^2}$$

which constitutes a parametric description of the curve.

Eliminating $t$ between the two equations (2), one gets an implicit equation that can be written under the form :

$$\tag{3} \left(x-\frac{6}{5}\right)^2+\left(y+\frac{3}{5}\right)^2=\frac{(y+2x)^2}{5}$$

Taking square roots on both sides:

$$\tag{4} \sqrt{\left(x-\frac{6}{5}\right)^2+\left(y+\frac{3}{5}\right)^2}=\frac{|y+2x|}{\sqrt{5}}$$

(4) expresses the fact that the distance of $M(x,y)$ (the current point on the curve) to point $F(\frac{6}{5},-\frac{3}{5})$ is equal to the distance of $M$ to straight line $(D)$ with equation $y+2x=0$ (with slope $-2$). For distance formula see: (http://mathworld.wolfram.com/Point-LineDistance2-Dimensional.html). Thus $F$ is the focus of the parabola and $(D)$ its directrix.

Therefore, the parabola's axis, passing through $F$ and orthogonal to $(D)$ (thus with slope $\frac{1}{2}$) has the following equation: $y+\frac{3}{5}=\frac{1}{2}(x-\frac{6}{5})$, i.e., $y=\frac{1}{2}x-\frac{6}{5}.$

Conclusion:

  • propositions (A), (B) are false. For (B), it's not necessary to compute the tangent at the vertex because the slope of this tangent is the same as the slope of the directrix, which is $-2$, not $\frac{1}{2}$.

  • propositions (C), (D) are exact.

enter image description here

Jean Marie
  • 81,803
1

Conditions (A),(B),(C),(D) are not needed to determine the tilted parabola ( $xy$ term non-zero). Because out of five constants needed to determine a conic, one can be reduced as zero determinant for parabola.You have given two points and two slopes which are quite sufficient.

$$ (x, y, y^{ \prime} ) = (3,3,1), (1,-1,-1) $$

Parabola involving $xy,x^2, y^2,x,y $ terms can be solved for y in a quadratic as:

$$ y^2 - 2y ( ax+b) + (ax+b)^2 - (cx+d) =0 $$

$$ y = (ax +b) \pm \sqrt{ c x +d } $$

$\pm$ symbol separates regions on either side of vertical tangent. Plug in given point coordinates:

$$ 3 a + b+ \sqrt{ 3c+d} =3 , \quad a + b+ \sqrt{ c+d} = -1 ,\quad a + c \ /( 2 \sqrt{ c+d} ) = 1 ,\quad a - c \ /( 2 \sqrt{ c+d} ) = -1 $$

Using a CAS to reduce tedium we get $ (a,b,c,d) = ( 1/2, -3/4, 9/4,-27/16) $

The parabola is plotted to verify everything .It checks out (C), (D) but (A),(B) are inconsistent with first inputs and are clearly incorrect.

EDIT1

Some calculations needed (in progress)at focus and tangent orthogonal intersection on directrix

For parabola $ 4 a y = x^2 $

Points of tangency $$ (2at, a t^2)\quad (2a/t, a/t^2) $$ Point of intersection of polar chord tangents on directrix which happens at right angles at a point D $$ a [( t-1/t), -1] $$ Length of tangents $T_1,T_2$ given by $$ (T_1/a)^2 = t^4+ 3 t^2 + 1/t^2 +3 = 2, \quad (T_2/a)^2 = 1/t^4+ 3/ t^2 + t^2 +3 = 18 $$

$$\rightarrow t= \frac13 $$

$OF$ is perpendicular on hypotenuse AB

$$ \frac{1}{OF^2} =\frac{1}{2} + \frac{1}{ 9 \cdot 2} $$

$$ OF = \frac{3}{\sqrt 5} ... $$

$$ \cos \beta= \frac{OF}{OB} = \frac{ 3}{\sqrt10},\quad \sin \beta= = \frac{ 1}{\sqrt10}\quad $$

to be continued

enter image description here

Narasimham
  • 40,495
  • I understand now: I should have had a closer eye. Besides, how can you say with your method that (C) (the focus is $(6/5,-3/5)$) is true ? Same remark for (D) ? – Jean Marie Oct 22 '16 at 10:44
  • You will think that I have something against you which is absolutely not the case, but it is matter of rigorous deduction. My concern is about your affirmation concerning the truth of propsition (C) by a rigorous deduction. The parabola you have found has, in Euclidean geometry, a very precise focus. Is it $(6/5,-3/5)$ or not, that is the question ? It is by no means an "over-determination" wrt to the defining data (2nd line of the question by the OP). – Jean Marie Oct 22 '16 at 12:10
  • I have nothing against you or anyone. Ici nous sommes tous amis.. I am in full agreement with your diagram and conclusions .We may go for a chat later on and co-operate with moderator. As far as (C) and (D) are concerned I observed purely by graphical appearance.that is all, I did not care to verify/check/prove etc. as that latter part was no more interesting to me. – Narasimham Oct 22 '16 at 13:05
  • @JeanMarie Sorry.You do have a point there. Shall try to answer the other correct part also.I can show the focal point after continuation of calculations done in my edit. – Narasimham Oct 22 '16 at 19:35
  • I appreciated your sentence in perfect French. Yes we are all here in a friendly spirit for helping young (or less young) people and sometimes arguing between us - this is healthy - especialy on my side for having a "scaled" degree of rigor, adapted as well to the level of the OPs. Sometimes, I use an introductory sentence "what follows is not rigorous, but it is intended to help the understanding..." – Jean Marie Oct 22 '16 at 20:29
  • Appreciate that. Initially I was bit put off as two propositions in a competition are false, made light of what was left. – Narasimham Oct 22 '16 at 20:37
1

It's well known that any rational quadratic Bézier curve is a conic section (ellipse, hyperbola, or parabola), and any (polynomial) quadratic Bézier curve is a parabola. There are fairly simple formulae for obtaining the geometric characteristics (directrix, focus, vertex) directly from the control points of the Bézier curve. For example, see the following paper and the earlier ones they cite:

Geometric Characteristics of Conics in Bézier Form
Cantóna, Fernández-Jambrina, Rosado María
Computer-Aided Design 43 (2011) 1413–1421

Here, we have a quadratic Bézier curve with control points $\mathbf{P}_0 =(3,3)$, $\mathbf{P}_1 = (0,0)$, $\mathbf{P}_2 = (1,-1)$. Several of the papers tell you that the axis of symmetry is in the direction of the vector $\tfrac12(\mathbf{P}_0 + \mathbf{P}_2) - \mathbf{P}_1$, which is $(2,1)$. Even without the papers, you can see this by letting $t\to\infty$ in the parametric equations of the curve. This shows that proposition (A) is false, and it follows immediately that (B) is also false.

Now that we know the axis direction, we can use the reflection property of a parabola to construct lines at $A$ and $B$ that pass through the focus. Intersecting these two lines gives us the focus. Or, use the formulae from the papers to get the focus.

I didn't write this up as a complete answer because I'm sure this is not how the problem was meant to be solved.

bubba
  • 43,483
  • 3
  • 61
  • 122