The Student Room Group

Prove f(x) can take all real values

I would appreciate a hint on how to approach this problem, which has so far resisted my attempts at its solution using the techniques described in the "Core Maths for A-Level" textbook I found it in.

"Provided that xx is real, prove that the function 2(3x+1)3(x29)\frac{2 (3x+1)}{3(x^2-9)} can take all real values."
The most obvious way at first glance would be to sketch the curve and hence explain that it takes all real values, as any horizontal line drawn will always intersect the curve. Alternatively, you can talk about what values it takes for certain intervals of x. For example for x < -3, f(x) takes all values less than 0; for x = 0, f(x) = 0; and finally, when x > 3, f(x) takes all values greater than 0 which again can be deduced from the graph, meaning that f(x) takes all real values. Things to consider whilst drawing are:

1. vertical asymptotes; think about what x makes the denominator 0.

2. turning points; differentiate and set to zero.

3. behaviour for large x; when x is a large positive number, f(x) approaches 0 from positive, whilst when x is a large negative number, f(x) approaches 0 from negative.

4. axis intercepts; set y and x = 0.

Of course this may not be the easiest approach, but thats what I would try if I were faced with something like this.
(edited 4 years ago)
Reply 2
You could restrict yourself to the interval (-3,3) for x.
Also note the if
y = 2(3x+1)/3(x^2-9)
can take on any value in the interval, so can
y = 3(x^2-9)/2(3x+1)
You can transform into a quadratic in x and show there is always a solution.

Not necessarily trivial and maybe not the way required.
(edited 4 years ago)
Original post by mqb2766
You could restrict yourself to the interval (-3,3) for x.
Also note the if
y = 2(3x+1)/3(x^2-9)
can take on any value in the interval, so can
y = 3(x^2-9)/2(3x+1)

Not quite true. Knowing that f(x) takes every value on (-inf, inf) doesn't guarantee that 1/f(x) ever takes the value 0.

Also, I don't think there's actually any advantage to flipping the RHS here: you want to show there are solutions for any choice of y, so the original equation becomes 3y(x^2-9) = 2(3x+1) which is still just a quadratic in x.

I don't know what "techniques described" the OP is talking about; I think the sketch you suggest is more "informative", but in an exam where I wanted to be sure of the marks, I think I'd go this algebra route.
Yeah just remembered from a-level, what you're meant to do is turn it into a quadratic in x, find the descrim and show that it is >=0 for all values of y :smile:.
Reply 5
Original post by DFranklin
Not quite true. Knowing that f(x) takes every value on (-inf, inf) doesn't guarantee that 1/f(x) ever takes the value 0.

Also, I don't think there's actually any advantage to flipping the RHS here: you want to show there are solutions for any choice of y, so the original equation becomes 3y(x^2-9) = 2(3x+1) which is still just a quadratic in x.

I don't know what "techniques described" the OP is talking about; I think the sketch you suggest is more "informative", but in an exam where I wanted to be sure of the marks, I think I'd go this algebra route.

Sure, the ppint at zero needs to be included (but its easy) and when i sketched the solution the flipped one looked a bit easier to prove, but as you say, they're both quadratics.
Reply 6
Original post by black1blade
Yeah just remembered from a-level, what you're meant to do is turn it into a quadratic in x, find the descrim and show that it is >=0 for all values of y :smile:.


Is this correct?


3y(x29)=2(3x+1)3y(x^2-9) = 2(3x+1) rearranges to 3yx26x(27y+2)=03yx^2-6x-(27y+2)=0

Finding the discriminant

324y2+24y+360324y^2+24y+36\geq0

Taking out a factor of 12

27y2+2y+3027y^2+2y+3\geq0

Completing the square

27(y2+1/27y)2+(80/27)>027(y^2+1/27y)^2+(80/27)>0

since

27(y2+1/27y)2027(y^2+1/27y)^2\geq0 for all real y, and 80/27>080/27>0
Reply 7
Original post by eljr
Is this correct?


3y(x29)=2(3x+1)3y(x^2-9) = 2(3x+1) rearranges to 3yx26x(27y+2)=03yx^2-6x-(27y+2)=0

Finding the discriminant

324y2+24y+360324y^2+24y+36\geq0

Taking out a factor of 12

27y2+2y+3027y^2+2y+3\geq0

Completing the square

27(y2+1/27y)2+(80/27)>027(y^2+1/27y)^2+(80/27)>0

since

27(y2+1/27y)2027(y^2+1/27y)^2\geq0 for all real y, and 80/27>080/27>0

Minor typo
27(y+1/27)2+(80/27)>027(y+1/27)^2+(80/27)>0
Yeah you shouldn't assume the descrim is always greater or equal to 0 til the last step. You could either complete the square to find the minimum of the quadratic descrim in y or differentiate. Normally you get a mark for during the graph as well to make it 100% clear that you know the descrim function is always greater than 0.

Quick Reply

Latest