0

I was looking for range of a rational function, I used three ways to get the answer and all of them are giving different results . e.g. f(x)=1/(1-x²) then,

(1)

 let 1/(1-x²) =y, (1-x²)=1/y

or

x²= 1-(1/y) or x= ±{1-(1/y)}½

Now 1-1/y must be ≥0

Therefore 1-1/y≥0 or 1≥1/y or y≥1

Therefore range (f)={1, ♾️}

(2)

let 1/(1-x²)=y then, 1= y-yx²

or (y)x²+0.x+(1-y)=0

Clearly D≥0 therefore -4y(1-y)≥0

Or

(i)

1-y≤0 or 1≤y 

this means range(f)={1,♾️}

(ii)

    -4y+4y²≥0 or y²≥y or y≥1
Again result is same as 2(i)

(3)

in f(x)= 1/(1-x²), putting

x= 0.0000001, y≈1

x≈ 0.999...., y≈ ♾️

x= 1.000...1, y≈ - ♾️

x= ±1000..., y≈ 0

Therefore

range (f)={(1, ♾️) U (-♾️, 0}

You can see that 1 and 2 yield same result but they are incorrect, I want to know that what is the mistake in logics there.

Note- I have rewritten the whole thing again after some of the comments below.

N. F. Taussig
  • 76,571
Level1
  • 47
  • 1
    Welcome to MSE. It is in your best interest that you type your posts (using MathJax) instead of posting links to pictures. – José Carlos Santos Jun 20 '22 at 06:35
  • This is my first-ever post in math community. I am not good at using MathJax. I have seen the topic how to use that but that looked time taking so I did it for ease. I may be able to use that slowly. – Level1 Jun 20 '22 at 06:40
  • Method 1 contains an obvious error when solving an inequality (when can you multiply by y?), method 2 contains a typo when rewriting an inequality (4 and y look similar, I guess), leading to an incorrect conclusion. Method 3 actually gives the right answer but is too informal to be considered a solution. Fixing the errors in 1,2 will give you the conclusion of 3. Also, there are many tools online to plot a function and see which answer is correct immediately on the plot. – Michal Adamaszek Jun 20 '22 at 06:45
  • Yes I saw that in 2nd method there is a typing error but even correcting that seems to yield range similar to the range obtained in 1st method, like I'm getting 4y²- 4y≥0 or y²-y≥0 or y²≥y or y≥1 again range {1, ♾️} – Level1 Jun 20 '22 at 07:01
  • No. $y^2\geq y$ is not equivalent to $y\geq 1$. This is exactly the same mistake you make in method 1. – Michal Adamaszek Jun 20 '22 at 07:04
  • But it is taught in inequalities that same operations can be done on both sides without affecting the inequality. – Level1 Jun 20 '22 at 07:08
  • 1
    $y^2\geq y$ is the same as $y^2-y\geq 0$, that is $y(y-1)\geq 0$, from which you see $y\in(-\infty,0]\cup[1,+\infty)$. I leave you to figure out yourself what happens when you use your method and divide both sides by $y$ (which can be positive or negative). You do the same in solution 1. – Michal Adamaszek Jun 20 '22 at 07:22
  • 1
    @user359206

    "But it is taught in inequalities that same operations can be done on both sides without affecting the inequality."

    This is not true. Multiplying an inequality with a negative number can most certainly not be done without affecting the inequality. Some operations can be done without affecting the inequality. This includes adding a value (i.e., $a<b$ is equivalent to $a+c<b+c$), and multiplying with a positive value (i.e., if $\alpha>0$, then $a<b$ is equivalent to $\alpha\cdot a < \alpha\cdot b$), but it does not include multiplying with an arbitrary value.

    – 5xum Jun 20 '22 at 07:29
  • That is why I added some operations, multiplying -ve reverses the inequality but that did not seem the case in above question. – Level1 Jun 20 '22 at 07:54
  • 2
    You seem to know then when an inequality is multiplied/divided by a negative number then the direction changes, because that's what you do in solution 2, where you have the number $-4$. But you somehow forget that knowledge when it comes to multiplying/dividing by $y$. Well, $y$ can be positive or negative so you need to consider those cases separately and either change the direction or not when multiplying by $y$. I mean, in solution 1 you agree that $y=-2$ satisfies the inequality $1\geq 1/y$? In that case, why doesn't your method find this solution? – Michal Adamaszek Jun 20 '22 at 07:58
  • Thank you for pointing out the mistake in 2nd method. But I am still looking for explanation for 1st method as why that approach is invalid. I can see that in 1-1/y≥0, putting y= -5 yields LHS ≥0 but solving 1-1/y≥0 gives y≥1. Where is the mistake in algebra? – Level1 Jun 20 '22 at 11:17

2 Answers2

0

We wish to find the range of the function $f(x) = \dfrac{1}{1 - x^2}$.

In what follows, I will assume that the range is a subset of the real numbers and that the domain is the largest subset of the real numbers for which the function is defined. Since the function is defined for every real number except those that make the denominator equal to zero, the domain is the set of all real numbers except $1, -1$. Thus, $$\text{Dom}_f = (-\infty, -1) \cup (-1, 1) \cup (1, \infty)$$

Method 1: You set $$y = \frac{1}{1 - x^2}$$ then took reciprocals. Before we can do that, we must first establish that $y \neq 0$.

If $y = 0$, we obtain $$\frac{1}{1 - x^2} = 0$$ Multiplying both sides of the equation by $1 - x^2$ yields $$1 = 0$$ which is a contradiction. Therefore, $y \neq 0$, so we can take reciprocals whenever $y$ is in the range.

\begin{align*} y & = \frac{1}{1 - x^2}\\ \frac{1}{y} & = 1 - x^2\\ x^2 & = 1 - \frac{1}{y} \end{align*} Thus, we require that $1 - \dfrac{1}{y} \geq 0$.

Observe that when $y < 0$, then $1 - \dfrac{1}{y} > 1 > 0$. You overlooked this case.

If $y > 0$, then \begin{align*} 1 - \frac{1}{y} & \geq 0\\ 1 & \geq \frac{1}{y}\\ y & \geq 1 \end{align*} where we have multiplied both sides of the inequality by $y > 0$ in the final step. Since $y > 0$, the direction of the inequality is preserved.

Hence, the range is $$\text{Ran}_f = (-\infty, 0) \cup [1, \infty)$$

Method 2: We know that if $y$ is in the range, then there is a real number $x$ such that $y = f(x)$. \begin{align*} y & = \frac{1}{1 - x^2}\\ y - yx^2 & = 1\\ 0 & = yx^2 + 1 - y \end{align*} which is a quadratic equation in $x$ provided that $y \neq 0$.

If $y = 0$, we obtain $0 = 1$, which is a contradiction. Hence, $y \neq 0$.

Since a quadratic equation has real roots whenever its discriminant $\Delta = b^2 - 4ac \geq 0$, there exists a real number $x$ such that $f(x) = y$ whenever \begin{align*} \Delta & \geq 0\\ b^2 - 4ac & \geq 0\\ 0^2 - 4y(1 - y) & \geq 0\\ 4y(y - 1) & \geq 0 \end{align*} and $y \neq 0$. The inequality holds when $y = 1$ or when the terms $y$ and $y - 1$ have the same sign. They have the same sign when $y < 0$ or $y > 1$. Hence, the range of the function is $$\text{Ran}_f = (-\infty, 0) \cup \{1\} \cup (1, \infty) = (-\infty, 0) \cup [1, \infty)$$ as we found above.

In both your attempts at this method, it looks like you divided by $y$, which caused you lose information. Rather than dividing by $y$, you should have factored the expression, as shown above.

N. F. Taussig
  • 76,571
  • If y<0, then 1-1/y≥0 then why have you not solve the inequality further as in case of y>0? – Level1 Jun 21 '22 at 06:48
  • Subtracting a negative number is equivalent to adding a positive number. If $y < 0$, $1/y < 0$. Therefore, $1/y$ from $1$ adds a positive number to $1$, so $1 - 1/y > 1$. – N. F. Taussig Jun 21 '22 at 09:04
0

Method 1:

let 1/(1-x²) =y, (1-x²)=1/y

so you should note that $y \ne 0$.

or

x²= 1-(1/y) or x= ±{1-(1/y)}½

Now 1-1/y must be ≥0

Therefore 1-1/y≥0 or 1≥1/y ...

So far so good but the next step is wrong

or y≥1

$1\ge \frac 1y \not \implies y\ge 1$

$1\ge \frac 1y \implies \begin{cases} y\ge 1&y> 0\\y\le 1&y < 0\end{cases}$

Therefore range (f)={1, ♾️}

Only if we assume $y$ is positive. If $y$ is negative then we have the only restriction being that $y \le 1$ but as $y < 0 \le 1$ that is redundant.

So the range is $(-\infty, 0)\cup [1,\infty)$.

....

Method 2:

let 1/(1-x²)=y then, 1= y-yx²

or (y)x²+0.x+(1-y)=0

Clearly D≥0 therefore -4y(1-y)≥0

So this means you have one of the following two cases.

Either

ONE: $1-y\ge 0$ AND $-4y\ge 0$.

(i)

1-y≤0 or 1≤y

You transposed the inequality symbol. You should have written

$1-y \ge 0$ or $1 \ge y$.

But You forgot to include $-4y \ge 0$. That will affect your answer:

$1-y\ge 0$ AND $-4y\ge 0$.

So $1 \ge y$ and $y\le \frac 0{-4} = 0$.

So $y \le 1$ AND $y \le 0$ so $y\le 0$. Combining that is $y \le 0 \le 1$. So $y \le 0$.

But $y = 0$ is impossible so $y < 0$ is part of the range.

Or

TWO: $1-y \le 0$ AND $-4y \le 0$

so $1\le y$ and $y \ge \frac 0{-4} = 0$. So $y \ge 1$ and $y \ge 0$. Combining we have $y\ge 1$.

So $[1,\infty)$ is the other part of the range.

Putting part ONE and TWO together we have either $y < 0$ or $y \ge 1$ so the range is $(-\infty, 0)\cup [1,-\infty)$.

-4y+4y²≥0 or y²≥y or y≥1

Again result is same as 2(i)

$y^2 \ge y \not \implies y \ge 1$.

It means if $y > 0$ then $y \ge 1$ but if $y=0$ you can not divide and if $y < 0$ then the inequality "flips".

So we have either.

a) $y > 0$ and $y\ge 1$ or

b) $y < 0$ and $y \le 1$ or

c) $y = 0$ but this is not possible.

So for the third time we get that the range is

$(-infty, 0) \cup [1, \infty)$.

All three methods give the same results.

fleablood
  • 124,253
  • From -4y(1-y)≥0, actually 1st time (i) I divided both sides by -4y to get rid of extra y, hence wrote 1-y≤0, second time (ii) I kept all y hence got -4y +4y²≥0, got rid of 4 and got -y+y²≥0 then y²≥y – Level1 Jun 21 '22 at 08:48
  • I thought any logical step or operation would be good but seems that I need more in-depth understanding. – Level1 Jun 21 '22 at 08:57
  • @user359206 Dividing by $y$ causes you to lose information. $y^2 \geq y$ whenever $y$ is negative. When you divided by $y$, you obtained $y \geq 1$, so you lost the information about negative solutions (and would have also lost the solution $y = 0$, which cannot occur in the original problem). Similarly, the sign of $-4y(1 - y)$ depends on both $y$ and $1 - y$, so you lose information when you divide by $y$. – N. F. Taussig Jun 21 '22 at 09:26
  • When I apply the same logic for (4-x²)^½, I get 4-x²=y² or x²+y²-4=0. Therefore -4(y²-4)≥0 or 4-y²≥0 or (2+y)(2-y)≥0 solving this I get y≥-2≥2 and y≤-2≤2 therefore y≥-2 and y≤2 But from domain it can be seen that range is [0,2]. Is the algebra misleading here? – Level1 Jul 14 '22 at 10:44
  • If $(4-x^2)^{\frac 12} \ge 0$. So if $y = (4-x^2)^{\frac 12}$ we know IMMEDIATELY that $y \ge 0$. So when you reach the conclusion that $y\ge -2$ and $y\le 2$ you combine that with $y\ge 0$ your final result is $0 \le y \le 2$. It is exactly the same..... when you square both sides, the square function is NOT one to one so again you have allowed extraneous possibilities (namely the possibility that $y$ might be negative when you KNOW it is not) to sneak in. – fleablood Jul 14 '22 at 16:12
  • " (2+y)(2-y)≥0 solving this I get y≥-2≥2 and y≤-2≤2" Huh? that's nonsense. $-2\not \ge 2$ so that makes no sense. If $(2+y)(2-y)\ge 0$ we have TWO different possibilities either 1) they are both greater or equal to $0$ *!!!!!!OR!!!!!!!* 2) they are both negative. If 1) $2+y \ge 0$ and $2-y\ge 0$ so $y\le -2$ and $y\ge 2$. That's contradictory so that is impossible. So we must have $2+y\ge 0$ and $2-y\ge 0$ so $y\le 2$ and $y\ge-2$ so $-2 \le y \le 2$. .... But as we ALWAYS knew $y \ge 0$ we have $0\le y \le 2$. – fleablood Jul 14 '22 at 16:21