Find the two points on the curve that share a tangent line

Click For Summary
The discussion focuses on finding two points on the curve y=x^4 - 2x^2 - x that share a tangent line, with the requirement of no calculators. Participants analyze the derivatives and set up equations based on the condition that the slopes at points (a, f(a)) and (b, f(b)) are equal. They derive relationships involving a and b, leading to the equations ab = 1/3 and a^2 + b^2 = 2/3. Confusion arises regarding the validity of solutions, particularly with the pairs (a= -1/sqrt(3), b= 1/sqrt(3)) and (a=-1, b=1), prompting discussions on checking the correctness of these solutions. Ultimately, the conversation highlights the complexity of the problem and the potential for simpler approaches to reach the solution.
Graxum
Messages
4
Reaction score
0
Homework Statement
Find the two points on the curve y=x^4 - 2x^2 - x that share a tangent line?
Relevant Equations
y=x^4 - 2x^2 - x

dy/dx = 4x^3 - 4x - 1
IMPORTANT: NO CALCULATORS

I assumed two points, (a, f(a)) and (b, f(b)) where b is greater than a. Since the tangent line is shared, I did

f'(a) = f'(b):

1) 4a^3 - 4a - 1 = 4b^3 - 4b - 1
2) 4a^3 - 4a = 4b^3 - 4b
3) 4(a^3 - a) = 4(b^3 - b)
4) a^3 - a = b^3 - b
5) a^3 - b^3 = a - b
6) (a - b)(a^2 + ab + b^2) = (a - b)
7) a^2 + ab + b^2 = 1

and

(f(b) - f(a))/b - a = f'(a):

1) (b^4 - 2b^2 - b - a^4 + 2a^2 + a) / (b - a) = 4a^3 - 4a - 1
2) [(b^4 - a^4) -2(b^2 - a^2) - (b - a)] / (b - a) = 4a^3 -4a -1
3) [(b^2 + a^2)(b + a)(b - a) - 2(b + a)(b - a) - (b - a)] / (b - a) = 4a^3 - 4a - 1
4) (b^2 + a^2)(b + a) - 2(b + a) - 1 = 4a^3 - 4a - 1
5) (b + a)(b^2 + a^2 - 2) = 4a(a - 1)(a + 1)
6) From the 7th step of the 1st equation, I deduced that:
a^2 + b(a + b) = 1
6.1) b(a + b) = (1 - a)(1 + a), thus -b(a + b) = (a - 1)(a + 1)
7) Subbing in -b(a + b) for (a - 1)(a + 1) on the right side of the equation on the 5th line, I got:
(b + a)(b^2 + a^2 - 2) = -4ab(a + b)
8) b^2 + a^2 - 2 = -4ab
9) From the 7th step of the 1st equation, I deduced that:
a^2 + b^2 = 1 - ab
10) Subbing in 1 - ab for a^2 + b^2 on the left side of the equation on the 8th line, I got:
-ab - 1 = -4ab
11) 1 = 3ab, thus ab = 1/3
12) Subbing in ab for 1 - a^2 - b^2 on the right side of the equation on the 8th line, I got:
a^2 + b^2 - 2 = -4 + 4a^2 + 4b^2
13) a^2 + b^2 = 2/3

Now you would think that knowing ab = 1/3 and a^2 + b^2 = 2/3 would make things easy, however I must've made a mistake somewhere because I find multiple possible solutions, a=1 (and f(a) for the y coordinate) and 2 values for b (those being 2/3 or -1, and their respective y coordinates.) I came so far and still can't find the answer!
 
Physics news on Phys.org
Graxum said:
Homework Statement:: Find the two points on the curve y=x^4 - 2x^2 - x that share a tangent line?
Relevant Equations:: y=x^4 - 2x^2 - x

dy/dx = 4x^3 - 4x - 1

IMPORTANT: NO CALCULATORS

I assumed two points, (a, f(a)) and (b, f(b)) where b is greater than a. Since the tangent line is shared, I did

f'(a) = f'(b):
...

and

(f(b) - f(a))/b - a = f'(a):
...

7) Subbing in -b(a + b) for (a - 1)(a + 1) on the right side of the equation on the 5th line, I got:
(b + a)(b^2 + a^2 - 2) = -4ab(a + b)
8) b^2 + a^2 - 2 = -4ab
...

Now you would think that knowing ab = 1/3 and a^2 + b^2 = 2/3 would make things easy, however I must've made a mistake somewhere because I find multiple possible solutions, a=1 (and f(a) for the y coordinate) and 2 values for b (those being 2/3 or -1, and their respective y coordinates.) I came so far and still can't find the answer!
Hello, @Graxum .

:welcome:

Be careful going from the result of step 7) to step 8).

You divided by ##(a+b)## . That is not legal if the value of ##(a+b)## is zero.

It's better to get zero on one side of the equation, then factor out ##(a+b)##.

That should also give the possibility that ##(a+b)=0##, in addition to the result you obtained.

Added in Edit:

At this point the problem is nearly solved. (Being a bit redundant, let's continue.)

##\displaystyle (b^2 + a^2 - 2 + 4ab)(a+b)=0##

So one of the two factors, ##\displaystyle (b^2 + a^2 - 2 + 4ab)## or ##\displaystyle (a+b)## must be zero.

You have already tried setting the first of these to zero. Now try the other, along with the 7th step of the 1st equation.
 
Last edited:
  • Like
Likes Graxum
Hint:
$$
\frac{dy}{dx}=\frac{y}{x}
$$
 
Fred Wright said:
Hint:
$$
\frac{dy}{dx}=\frac{y}{x}
$$
That doesn't seem to me to be correct.
 
Graxum said:
(f(b) - f(a))/b - a = f'(a):
I think keeping the problem symmetric in ##a## and ##b## will significantly simplify your algebra. Try writing your equation for the shared tangent line as:$$\frac{f(b)-f(a)}{b-a}=\frac{f^{\prime}(a)+f^{\prime}(b)}{2}$$ and repeating your derivation.
 
Let the tangent line be y = mx + c, and define <br /> g(x) = f(x) - mx - c = x^4 - 2x^2 - (m+1)x - c. Npw if (a, f(a)) is on the tangent line then g(a) = 0, and for the line to actually be a tangent we require g&#039;(a) = 0. We obtain the same conditions for b. Can you see why the fact that g is a polynomial of order 4 in x then requires <br /> g(x) = (x- a)^2(x- b)^2 = x^4 - 2x^2 - (m+1)x - c? Comparing coefficients of x^3 and x^2 will now give you an easier system of equations to solve for a and b.
 
  • Like
Likes SammyS
SammyS said:
Hello, @Graxum .

:welcome:

Be careful going from the result of step 7) to step 8).

You divided by ##(a+b)## . That is not legal if the value of ##(a+b)## is zero.

It's better to get zero on one side of the equation, then factor out ##(a+b)##.

That should also give the possibility that ##(a+b)=0##, in addition to the result you obtained.

Added in Edit:

At this point the problem is nearly solved. (Being a bit redundant, let's continue.)

##\displaystyle (b^2 + a^2 - 2 + 4ab)(a+b)=0##

So one of the two factors, ##\displaystyle (b^2 + a^2 - 2 + 4ab)## or ##\displaystyle (a+b)## must be zero.

You have already tried setting the first of these to zero. Now try the other, along with the 7th step of the 1st equation.
Ah i completely forgot about that!

I have 1 more question though, After the correction I got 2 answers ##(a= -1/sqrt3, b= 1/sqrt3)## from the left side and ##(a=-1, b=1)## from the right side, how do I check which of those is the correct answer? are both correct? in that case, when I plot the main equation on a graph, the 1st one seems to be completely off and those points definitely don't share a tangent line.

I wonder if I calculated something wrong somewhere?
 
Graxum said:
Ah i completely forgot about that!

I have 1 more question though, After the correction I got 2 answers ##(a= -1/\sqrt 3, b= 1/\sqrt 3)## from the left side and ##(a=-1, b=1)## from the right side, how do I check which of those is the correct answer? are both correct? in that case, when I plot the main equation on a graph, the 1st one seems to be completely off and those points definitely don't share a tangent line.

I wonder if I calculated something wrong somewhere?
Well, you could plug those values into check them.

Beyond that:
For the following equation to be true,

##\displaystyle (b^2 + a^2 - 2 + 4ab)(a+b)=0##

all that's required is that either ##\displaystyle (b^2 + a^2 - 2 + 4ab)=0 \text{ or } (a+b)=0##

It looks like you found that ##a+b=0## leads to ##a=-1, \ b=1## .

That does give the correct solution.
 
I thought I was fairly well up in polynomials, for an amateur, and that this would be a doddle. Instead I found it quite difficult, maybe the operative word there is 'was'. :oldcry: .

I didn't go through the OP's calculations, they looked too heavy and complicated. When it gets that complicated one should look for a simpler approach.

But then I realize that is even when true easier said than done sometimes. especially if you had a method which you are sure should work, so if it doesn't you might think there is some calculational error and ask for help as the OP did.

I know how to find the condition for a double root of a polynomial, I do not know a formulation for finding two double roots any more compact than finding one and then the other.

Anyway I eventually found (independently) essentially the same formulation as that of pasmith, and it all falls out in a couple of lines, with result ##m = -1, c=-1##. You do not even need to consider the derivative. So I can only add, seeing is believing:

31C40C5F-AA26-4C21-8E98-068A3D02F63A.jpeg


I'll worry about how general this is some other time.
 
Last edited: