MHB Max Value of |bc| for Complex Numbers in Inequality

  • Thread starter Thread starter anemone
  • Start date Start date
  • Tags Tags
    Maximum
anemone
Gold Member
MHB
POTW Director
Messages
3,851
Reaction score
115
$a,\,b,\,c$ are the complex numbers in $|az^2+bz+c| \le 1$ for any complex number $z$ where $|z| \le 1$. Find the maximum value of $|bc|$.
 
Mathematics news on Phys.org
If $|b|$ and $|c|$ are to be as large as possible, it looks likely that $|a|$ will have to be small. So a first attempt would be to see what happens if $a=0$. In that case, the function becomes $|bz+c|$. That function takes the unit disc to a disc of radius $|b|$ with its centre at a distance $|c|$ from the origin. The condition for that image to lie in the unit disc is $|b| + |c| \leqslant 1$. The maximum value of $|bc|$ is then $\frac14$, occurring when $|b|$ and $|c|$ are both $\frac12$.

So my first guess was that the answer to the problem might be $\frac14$. But I then did some numerical experiments with graphs, and I found that if $a = 0.16$, $b = 0.56$ and $c = -0.56$, then the function $0.16z^2 + 0.56z - 0.56$ takes the unit disc to the green region in the diagram below, which clearly lies inside the (red) unit circle. But it has the property that $|bc| = 0.56^2 = 0.3136$, which is quite a bit larger than $\frac14$.

That is as far as I can go with this problem. I don't even have a guess as to what the maximum value of $|bc|$ might be.

[TIKZ][scale=6]\draw [help lines, ->] (-1.1,0) -- (1.1,0) ;
\draw [help lines, ->] (0,-1.1) -- (0,1.1) ;
\draw [red] circle (1) ;
\fill [green, domain=0:6.285, samples=100] plot ({0.16*cos(2*\x r) + 0.56*cos(\x r) - 0.56}, {0.16*sin(2*\x r) + 0.56*sin(\x r)});[/TIKZ]
 
This is a solution for the case where the parameters $a,b,c$ are all real. I suspect that the answer may be the same if they are allowed to be complex, but the proof would have to be quite different.

For the condition $|az^2 + bz + c| \leqslant 1$ to hold whenever $|z|\leqslant1$, it is sufficient (by the maximum modulus principle) to check that it holds when $|z|=1$. So we want to ensure that $|ae^{2i\theta} + be^{i\theta} + c| \leq 1$ (for all $\theta$). But if $a,b,c$ are real then $$\begin{aligned}|ae^{2i\theta} + be^{i\theta} + c|^2 &= (ae^{2i\theta} + be^{i\theta} + c)(ae^{-2i\theta} + be^{-i\theta} + c) \\ &= 2ac\cos(2\theta) + 2b(a+c)\cos\theta + a^2 + b^2 + c^2 \\ &= 2ac(2\cos^2\theta - 1) + 2b(a+c)\cos\theta + a^2 + b^2 + c^2 \\ &= 4acx^2 + 2b(a+c)x + (a-c)^2 + b^2, \text{ where }x = \cos\theta. \end{aligned}$$ So we want $f(x) = 4acx^2 + 2b(a+c)x + (a-c)^2 + b^2$ to have maximum value $1$ on the interval $-1\leqslant x \leqslant 1$. Differentiate to get $f'(x) = 8acx + 2b(a+c)$. That is zero when $x = -\dfrac{b(a+c)}{4ac}$, and at that point $$f(x) = \frac{b^2(a+c)^2}{4ac} - \frac{2b^2(a+c)^2}{4ac} + (a-c)^2 + b^2.$$ Put that equal to $1$ and solve for $b^2$, to get $$b^2 = 4ac\left(1 - \frac1{(a-c)^2}\right).$$ But for fixed $c$ we want to choose $a$ so as to maximise $b$. So differentiate $b^2$ with respect to $a$, getting $$4c\left(1 - \frac1{(a-c)^2}\right) + \frac{8ac}{(a-c)^3}.$$ Put that equal to zero to get $(a-c)^3 = -(a+c).$

Now let $t = a-c$. Then $t^3 = -a-c$, so that $c = \frac12(t^3 + t)$. Also, $t^6 - t^2 = (a+c)^2 - (a-c)^2 = 4ac$. Therefore $$b^2c^2 = (t^6 - t^2)\left(1 - \frac1{t^2}\right)\cdot\frac14(t^3+t)^2 = \frac14t^2(t^2-1)^2(t^2+1)^3.$$ On the interval $-1\leqslant t\leqslant 1$, that function has a maximum value $\dfrac{27}{256}$ (attained when $t =\pm \dfrac1{\sqrt2}$). So the maximum value of $|bc|$ is $\dfrac{3\sqrt3}{16} \approx 0.3248.$

From that, it is easy to see that this maximum value occurs when $a = \dfrac{\sqrt2}8 \approx 0.177$, $b = \dfrac{\sqrt6}4 \approx 0.612$ and $c =
-\dfrac{3\sqrt2}8 \approx -0.53$. Using those numbers, the diagram from the previous comment looks like this, with the green image of the unit disc under the function $|az^2 + bz + c|$ fitting snugly inside the unit circle.
\begin{tikzpicture} [scale=6]\draw [help lines, ->] (-1.1,0) -- (1.1,0) ; \draw [help lines, ->] (0,-1.1) -- (0,1.1) ; \draw [red] circle (1) ; \fill [green, domain=0:6.285, samples=100] plot ({0.177*cos(2*\x r) + 0.612*cos(\x r) - 0.53}, {0.177*sin(2*\x r) + 0.612*sin(\x r)}); \end{tikzpicture}
 
Insights auto threads is broken atm, so I'm manually creating these for new Insight articles. In Dirac’s Principles of Quantum Mechanics published in 1930 he introduced a “convenient notation” he referred to as a “delta function” which he treated as a continuum analog to the discrete Kronecker delta. The Kronecker delta is simply the indexed components of the identity operator in matrix algebra Source: https://www.physicsforums.com/insights/what-exactly-is-diracs-delta-function/ by...
Fermat's Last Theorem has long been one of the most famous mathematical problems, and is now one of the most famous theorems. It simply states that the equation $$ a^n+b^n=c^n $$ has no solutions with positive integers if ##n>2.## It was named after Pierre de Fermat (1607-1665). The problem itself stems from the book Arithmetica by Diophantus of Alexandria. It gained popularity because Fermat noted in his copy "Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos, et...
Thread 'Imaginary Pythagorus'
I posted this in the Lame Math thread, but it's got me thinking. Is there any validity to this? Or is it really just a mathematical trick? Naively, I see that i2 + plus 12 does equal zero2. But does this have a meaning? I know one can treat the imaginary number line as just another axis like the reals, but does that mean this does represent a triangle in the complex plane with a hypotenuse of length zero? Ibix offered a rendering of the diagram using what I assume is matrix* notation...

Similar threads

Replies
1
Views
1K
Replies
1
Views
1K
Replies
2
Views
1K
Replies
7
Views
2K
Replies
2
Views
1K
Replies
7
Views
2K
Replies
1
Views
1K
Back
Top