1

How to find the sum of the complex roots of the quadratic equation $$px^2+qx+r=0$$ if $r>p>0$, and $p,q,r \in\mathbb R$ ?

Lucian
  • 48,334
  • 2
  • 83
  • 154

2 Answers2

1

The given equation is

$px^2+qx+r=0, \; \text{with} \; r > p > 0 \tag{1}$

and $p, q, r \in \Bbb R$, the real numbers. If $\rho_1$ and $\rho_2$ are the roots of (1), then they are also roots of

$x^2+\dfrac{q}{p}x+\dfrac{r}{p} = 0, \tag{2}$

which is just (1) divided through by $p \ne 0$. $\rho_1$ and $\rho_2$ being roots of (2) implies we have the factorization

$x^2+\dfrac{q}{p}x+\dfrac{r}{p} = (x - \rho_1)(x - \rho_2) = x^2 - (\rho_1 + \rho_2)x + \rho_1 \rho_2, \tag{3}$

and if we compare the coefficients of like powers of $x$, i.e., of $x^0 = 1, x^1 = x$ and $x^2$, we see that

$\rho_1 + \rho_2 = -\dfrac{q}{p} \tag{4}$

and

$\rho_1 \rho_2 = \dfrac{r}{p}. \tag{5}$

(4) and (5) hold in the both the cases real and complex roots. In fact, it is easy to see that they hold over any field, i.e. when $0 \ne p, q, r \in \Bbb F$ an arbitrary field in which (1) has a root.

Note Added in Edit: Returning to the case $\Bbb F = \Bbb R$, $r >p > 0$, with $\rho_1, \rho_2 \in \Bbb C$ but $\rho_1, \rho_2 \notin \Bbb R$, note that in this event we have $\rho_2 = \bar {\rho}_1$, i.e. $\rho_1$ and $\rho_2$ are complex conjugates.Thus by (5),

$\vert \rho_1 \vert^2 = \rho_1 \bar{\rho_1} = \rho_1 \rho_2 = \dfrac{r}{p}, \tag{6}$

and likewise

$\vert \rho_2 \vert^2 = \dfrac{r}{p}, \tag{7}$

so that

$\vert \rho_1 \vert = \vert \rho_2 \vert = \sqrt{\dfrac{r}{p}}. \tag{8}$

Thus

$\vert \rho_1 \vert + \vert \rho_2 \vert = 2\sqrt{\dfrac{r}{p}} > 2 \tag{9}$

since $r > p$. This is option (d) in user140013's comment. End of Note.

Hope this helps. Cheerio,

and as always,

Fiat Lux!!!

Robert Lewis
  • 71,180
0

$$ x = \frac{-q \pm \sqrt{q^2 - 4pr}}{2p} $$

Regardless:

$$ \frac{-q + \sqrt{q^2 - 4pr}}{2p} + \frac{-q - \sqrt{q^2 - 4pr}}{2p} =\frac{-2q + \sqrt{q^2 - 4pr} - \sqrt{q^2 - 4pr}}{2p}= \frac{-2q + 0}{2p} = -\frac{q}{p} $$

edit

OK, I think I understand the problem now. For there to be complex roots, we have:

$$ q^2 < 4pr \rightarrow |q| < 2\sqrt{pr} $$

We need to handle two different cases:

  1. q > 0
  2. Here we have $-\frac{q}{p} = -\frac{|q|}{p}$. Since $|q| < 2\sqrt{pr}$, this gives a bound from a "small" negative number. The actual $q$ must be smaller, and thus must be a smaller negative which is really a larger number (a more positive number), thus we get: $$ \text{sum} > -2\frac{\sqrt{pr}}{p}\rightarrow \text{sum} > -2\sqrt{\frac{r}{p}} $$ Since $r > p$, $\sqrt{\frac{r}{p}} > 1$. This gives a minimum (negative) number of $-2$. Since the sum should be greater, this means we indeed have $\text{sum} > -2$.
  3. $q \leq 0$
  4. In this case $-\frac{q}{p} = \frac{|q|}{p}$. Now we need a legitimately smaller number (smaller positive number): $$ \text{sum} < 2\sqrt{\frac{r}{p}} < 2 $$

This would suggest that the sum is between $-2$ and $+2$.

Jared
  • 6,227
  • But choices to this question are as : a)equal to 1 b)equal to 2 c)less than 2 but not equal to 1 d)greater than 2 – user140013 Apr 03 '14 at 04:32
  • The sum of the complex roots could be $0$ if $q^2 \geq 4pr$. I don't see anything in the question that requires this though. – Jared Apr 03 '14 at 04:36
  • Sorry I gave you wrong question mod(one root)+mod(other root) is asked – user140013 Apr 03 '14 at 04:39
  • 1
    @user140013: you might consider editing your question if it is wrong as stated. Wrong meaning, not what you (apparently) intended, judging from your comment. – Robert Lewis Apr 03 '14 at 05:00
  • Yeah, and I'm still questioning my reasoning on what I assumed to be the question. I agree though, the question needs to be clarified (a lot). The sum could be less than $-2$. For instance $r = 4p$ and $q < 4p$. So maybe $p = 1, r = 4, q = 3$. This would give $-\frac{q}{p} = -\frac{3}{1} < -2$. – Jared Apr 03 '14 at 05:05
  • 1
    @user140013: what does mod(one root) + mod(other root) mean? I'm guessing you mean $\vert \rho_1 \vert + \vert \rho_2 \vert$, where the $\rho_i$ are the roots of $px^2 + qx + r = 0$. A very different question indeed! Now I have to rethink and perhaps edit my answer. Which is, of course, OK! – Robert Lewis Apr 03 '14 at 05:06
  • I think I'm going to delete my answer because even for what I assumed, it's not correct. The sum could be anywhere from $(-\infty, +\infty)$ (even under the constraint that we definitely have two complex roots)...because $p$ could be very small, $r$ very large and thus $q$ very large. – Jared Apr 03 '14 at 05:09
  • I believe the sum of the magnitudes of the two complex roots would just be $\frac{1}{p}\sqrt{2q^2 - 4pr}$ under the constraint that $q^2 < 4pr$ which, I think, would give a bound $\frac{\sqrt{2\cdot 4pr - 4pr}}{p} = \frac{2\sqrt{pr}}{p} = 2\sqrt{\frac{r}{p}}$. – Jared Apr 03 '14 at 05:14
  • ...but that's really no bound at all since $\sqrt{\frac{r}{p}} > 1$ means it could be very large and since that was the upper bound, the sum of the magnitudes just has to be smaller than some (possibly) unbounded value. – Jared Apr 03 '14 at 05:22
  • @user140013: my answer explains why $\vert \rho_1 \vert + \vert \rho_2 \vert > 2$. So choice (d) is correct. – Robert Lewis Apr 03 '14 at 05:39
  • @Jared: you might want to read my answer to this question. Cheers! – Robert Lewis Apr 03 '14 at 05:39