I need to fit a parabola between

1) $y = -0.25x + 1.5$, on the interval $[0, 2]$

2) $y = 3$, on the interval $[4,6]$

Such that the resulting curve is continuous.

My thought process was as follows: Since it only has to be continuous, I need to make sure the right- and left endpoint of the graphs need to be hit.

I thought to fit a parabola using $(2,1)$ as the vertex of the parabola and making sure $(4, 3)$ (since that's the right endpoint of $y=3$), is on the graph.

As a possible solution, I came up with the parabola: $0.5x^2 - 2x + 3$ (see figure below). I would argue that $$\lim_{x \to 4^-} 0.5x^2 - 2x + 3 == \lim_{x \to 4^+} 3$$

Paint-skilled some intervals

Alas, my solution was incorrect. What would be a better way to approach this?


The required parabola is $$\color{red}{y=0.625x^2-2.75x+4}$$

Both $(2,1)$ and $(4,3)$ lie on this parabola, and the slope at $(2,1)$ is $-0.25$.

The curve is obtained by using a general quadratic and specifying that both the given points lie on it, and that the slope $2ax+b$ at $(2,1)$ is $-0.25$.

enter image description here

  • 1
    $\begingroup$ This indeed is the correct answer $\endgroup$ – Apeiron Jun 28 '17 at 10:22
  • $\begingroup$ ?? it's not differentiable at x=4 $\endgroup$ – Saketh Malyala Jun 29 '17 at 23:17
  • $\begingroup$ From the diagram in the question it appears that OP wants it to be differentiable at $(2,1)$. Not possible to be both. $\endgroup$ – hypergeometric Jun 30 '17 at 0:36
  • $\begingroup$ @Apeiron - Thank you. Glad to know it was useful. $\endgroup$ – hypergeometric Jun 30 '17 at 16:08

You actually almost had it. Probably some silly mistake.

We place the vertex at $(2,1)$, and make sure it hits the point $(4,3)$.

Our parabola should be of the form $y=a(x-2)^2+1$.

Plugging in $y=3$, $x=4$, gives us $3=a(4-2)^2+1$, or $3=4a+1$, so $\displaystyle a=\frac{1}{2}$.

We have $\displaystyle y=\frac{1}{2}(x-2)^2+1$.

If you want to put that into standard form, it is $\displaystyle y=\frac{1}{2}x^2-2x+3$.

Which is what you got.

This curve is continuous. SO how exactly are you wrong? enter image description here


The curve must be differentiable.

We have the curve $y=ax^2+bx+c$.

We obtain $y'=2ax+b$.

We need that $y'(4)=0$, so $0=8a+b$.

We also need that $y'(2)=-0.25$, so $-0.25=4a+b$.

Solving this system gives us $\displaystyle a=\frac{1}{16}$, and $\displaystyle b=-\frac{1}{2}.$

Therefore, $\displaystyle y=\frac{1}{16}x^2-\frac{1}{2}x+c$.

We plug in $y=3$, $x=4$, to get $3=1-2+c$, so $c=4$.


Just kidding. This isn't even possible.

By the Mean Value Theorem, we need some point on the parabola to have slope $1$.

A fitted parabola must begin decreasing, then start increasing. However, if it begins increasing, then it can no longer have a flat tangent at $x=4$, and be differentiable to $y=3$.

  • $\begingroup$ Hmm, after reading the exercise again, I should have made sure the curve is smooth i.e. tangent lines of endpoints must match (so it does have to be differentiable). And: what software do you use to graph? $\endgroup$ – Apeiron Jun 28 '17 at 7:55
  • 1
    $\begingroup$ I used desmos. also in that case i'll show you how to make it differentiable $\endgroup$ – Saketh Malyala Jun 28 '17 at 8:00
  • $\begingroup$ The answer is not correct. $y=\frac {x^2}{16}-\frac x2+4$ does not pass through $(2,1)$. $\endgroup$ – hypergeometric Jun 28 '17 at 13:50
  • $\begingroup$ The parabola cannot be differentiable and pass through both points. @hypergeometric $\endgroup$ – Saketh Malyala Jun 29 '17 at 19:27
  • $\begingroup$ It can pass through both points and be differentiable at one, in this case $(2,1)$. $\endgroup$ – hypergeometric Jun 30 '17 at 0:37

An endpoint is (2; 1) on the slant line. Another endpoint is (4; 3) on the horizontal line. So consider a general parabola of equation


We know that for $x=2$ must be $y=1$ and for $x=4$ must be $y=3$. Thus substituting you get two equations

$$4 a + 2 b + c=1;\;16 a + 4 b + c = 3$$

Solving for $a;\;b$ you get

$$a= \frac{c+1}{8};\;b= \frac{1}{4} (1-3 c)$$

if you want no fraction you can freely set $c=7$ and have the parabola

$$y=x^2 - 5 x + 7$$

Hope it is clear

  • $\begingroup$ Nope. It isn't differentiable. $\endgroup$ – Saketh Malyala Jun 29 '17 at 19:27

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.