Hard complex number question (1 Viewer)

cutemouse · Jul 22, 2009

Hi,

Could someone please help me out here?

Thanks

kaz1 · Jul 22, 2009

For the first one try taking out 2ⁿsecⁿ@

Trebla · Jul 22, 2009

For (i), write tan θ = sin θ / cos θ and factor out everything until you end up with cis θ so that De Moivre's theorem can be applied.

For (ii), from (i) let θ = π/10 and n = 5, which gives:
(2 + 2i tan π/10)⁵ + (2 - 2i tan π/10)⁵ = (2⁶cos π/2) / (cos π/10)⁵
The RHS = 0, hence
(2 + 2i tan π/10)⁵ + (2 - 2i tan π/10)⁵ = 0
=> z = 2i tan π/10 satisfies the equation:
(2 + z)⁵ + (2 - z)⁵ = 0

For (iii), the roots take the form 2i tan θ, so we solve cos 5θ = 0 from the RHS of (i), choose the solutions of θ that satisfy this trigonometric equation and since your roots are of the form, 2i tan θ you can then list all 5 roots explicitly given the values of θ found (NB: that you cannot choose 5π/10 since this would make 2i tan θ undefined).

For (iv), expand (ii) using binomial theorem and equate the real parts and let z = 2i tan π/10. This should give a quadratic equation which can be solved for (tan π/10)²

jet · Jul 22, 2009

$LHS = \left (2 + 2itan\, \theta \right )^n + \left (2 - 2itan\, \theta \right )^n \\ = 2^n\left (\frac{cos\theta + isin\theta}{cos\theta} \right )^n +2^n\left (\frac{cos\theta - isin\theta}{cos\theta} \right )^n \\ = \frac{2^n}{cos^n\theta} \left [\left (cos\theta + isin\theta \right )^n + \left (cos\theta - isin\theta \right )^n \right ] \\ = \frac{2^n}{cos^n\theta} \left [\left (cos \, n\theta + isin \, n\theta \right ) + \left (cos \, n\theta - isin \, n\theta \right ) \right ] \text{ noting that } cos \, n\theta - isin \, n\theta = cos \left (-\theta \right ) + isin \left (-\theta \right ) \\ = \frac{2^n}{cos^n\theta} \left (2cos \, n\theta \right ) \\ = \frac{2^{n+1}cos \, n\theta}{cos^n \, \theta} \\ = RHS$

$\text{Letting } z = 2i tan \, \theta, \\ \left (2 + z \right )^n + \left (2 - z \right )^n = \frac{2^{n+1}cos \, n\theta}{cos^n \, \theta} \\ \text{Thus, to solve } \left (2 + z \right )^5 + \left (2 - z \right )^5 = 0, \text{ we can solve:} \\ \frac{2^{6}cos \, 5\theta}{cos^5 \, \theta} = 0 \\ cos \, 5\theta = 0 \\ 5\theta = \frac{\pi}{2}, \frac{-\pi}{2}, \frac{3\pi}{2}, \frac{-3\pi}{2}, \text{ etc.} \\ \theta = \frac{\pi}{10}, \text{ etc.} \\ \text{Thus, one solution is } \theta = \frac{\pi}{10}, \text{ which corresponds to the root } z = 2itan \, \frac{\pi}{10} \\ \text{Hence, } 2itan \, \frac{\pi}{10} \text{ is a root of the equation } \left (2 + z \right )^5 + \left (2 - z \right )^5 = 0 \\$

$\text{Taking } 0 \leq \theta \leq \pi \text{ since the period of tan is } \pi, \text{the solutions of the equation are } \theta = \frac{\pi}{10}, \frac{3\pi}{10},\frac{5\pi}{10}, \frac{7\pi}{10}, \frac{9\pi}{10}\\ \text{Hence the other roots are } z = 2itan \, \frac{m\pi}{10}, \text{where} m = 1,3,5,7,9$

Will do last part after dinner.

jet · Jul 22, 2009

cutemouse · Jul 22, 2009

Thanks jetblack, it makes alot of sense to me now.

But 1 question though... Where did you get the tan(pi/10) < tan(pi/4) ?

Thanks

untouchablecuz · Jul 22, 2009

jetblack2007 said:
$\left (2 + z \right )^5 + \left (2 - z \right )^5 = 0 \\ 32 + 80z + 80z^2 + 40z^3 + 10z^4 + z^5 + 32 - 80z + 80z^2 - 40z^3 + 10z^4 - z^5 = 0 \\ 64 + 160z^2 + 20z^4 = 0 \\ \text{Since } z = 2itan \, \frac{\pi}{10} \text{ is a root} \\ 320tan^4 \frac{\pi}{10} - 640tan^2 \frac{\pi}{10} + 64 = 0 \\ 5tan^4 \frac{\pi}{10} - 10tan^2 \frac{\pi}{10} + 1 = 0 \\ \therefore tan^2 \frac{\pi}{10} = \frac{10 \pm \sqrt{100 - 4(1)(5)}}{10} \\ = \frac{10 \pm \sqrt{80}}{10} \\ = 1 + \frac{2\sqrt{5}}{5} \text{ or } 1 - \frac{2\sqrt{5}}{5} \\ \text{But, since } tan\frac{\pi}{10} < tan \frac{\pi}{4} = 1, \\ tan^2 \frac{\pi}{10} = 1 - \frac{2\sqrt{5}}{5}$

just one Q, with this question, did you know to expand the binomials (evetnually leading to a quadratic in x^2) or just tried it in an attempt to answer the Q?

untouchablecuz · Jul 22, 2009

jm01 said:
Thanks jetblack, it makes alot of sense to me now.

But 1 question though... Where did you get the tan(pi/10) < tan(pi/4) ?

Thanks

think of the graph of y = tan x

Aerath · Jul 22, 2009

jm01 said:
But 1 question though... Where did you get the tan(pi/10) < tan(pi/4) ?

I believe you can just say: "by inspection". Or just say: "using calculator".

gurmies · Jul 22, 2009

1/10 < 1/4

For both sinx and tanx as x ===> pi/2, sinx and tanx ===> 1... for cosx it approaches zero. [0<=x<=pi/2]

Trebla · Jul 22, 2009

jetblack2007 said:
$LHS = \left (2 + 2itan\, \theta \right )^n + \left (2 - 2itan\, \theta \right )^n \\ = 2^n\left (\frac{cos\theta + isin\theta}{cos\theta} \right )^n +2^n\left (\frac{cos\theta - isin\theta}{cos\theta} \right )^n \\ = \frac{2^n}{cos^n\theta} \left [\left (cos\theta + isin\theta \right )^n + \left (cos\theta - isin\theta \right )^n \right ] \\ = \frac{2^n}{cos^n\theta} \left [\left (cos \, n\theta + isin \, n\theta \right ) + \left (cos \, n\theta - isin \, n\theta \right ) \right ] \text{ noting that } cos \, n\theta - isin \, n\theta = cos \left (-\theta \right ) + isin \left (-\theta \right ) \\ = \frac{2^n}{cos^n\theta} \left (2cos \, n\theta \right ) \\ = \frac{2^{n+1}cos \, n\theta}{cos^n \, \theta} \\ = RHS$

$\text{Letting } z = 2i tan \, \theta, \\ \left (2 + z \right )^n + \left (2 - z \right )^n = \frac{2^{n+1}cos \, n\theta}{cos^n \, \theta} \\ \text{Thus, to solve } \left (2 + z \right )^5 + \left (2 - z \right )^5 = 0, \text{ we can solve:} \\ \frac{2^{6}cos \, 5\theta}{cos^5 \, \theta} = 0 \\ cos \, 5\theta = 0 \\ 5\theta = \frac{\pi}{2}, \frac{-\pi}{2}, \frac{3\pi}{2}, \frac{-3\pi}{2}, \text{ etc.} \\ \theta = \frac{\pi}{10}, \text{ etc.} \\ \text{Thus, one solution is } \theta = \frac{\pi}{10}, \text{ which corresponds to the root } z = 2itan \, \frac{\pi}{10} \\ \text{Hence, } 2itan \, \frac{\pi}{10} \text{ is a root of the equation } \left (2 + z \right )^5 + \left (2 - z \right )^5 = 0 \\$

$\text{Taking } 0 \leq \theta \leq \pi \text{ since the period of tan is } \pi, \text{the solutions of the equation are } \theta = \frac{\pi}{10}, \frac{3\pi}{10},\frac{5\pi}{10}, \frac{7\pi}{10}, \frac{9\pi}{10}\\ \text{Hence the other roots are } z = 2itan \, \frac{m\pi}{10}, \text{where} m = 1,3,5,7,9$

Will do last part after dinner.

Just to make note, there is a problem at m = 5 since z = 2i tan 5π/10 is undefined.

jet · Jul 22, 2009

Trebla said:
Just to make note, there is a problem at m = 5 since z = 2i tan 5π/10 is undefined.

That's interesting. I'm just unsure how to rectify it....

EDIT: Thinking over it, you can see that z^5 cancels out in the equations, meaning that it is actually a polynomial of degree 4 = four roots. I'm thinking just skip m=5.

jet · Jul 22, 2009

untouchablecuz said:
just one Q, with this question, did you know to expand the binomials (evetnually leading to a quadratic in x^2) or just tried it in an attempt to answer the Q?

Well from the fact that it contained tan pi/10, it was obvious that the root from one of the previous parts came into it. I just worked from there.

untouchablecuz · Jul 22, 2009

jetblack2007 said:
Well from the fact that it contained tan pi/10, it was obvious that the root from one of the previous parts came into it. I just worked from there.

well i realised that part, but did u know that um er

i dunno how to explain myself, dw

jet · Jul 22, 2009

untouchablecuz said:
well i realised that part, but did u know that um er

i dunno how to explain myself, dw

Well there was really nothing else to do.

cutemouse · Jul 22, 2009

Thanks alot Jetblack. I really understand it now =D

cutemouse · Jul 22, 2009

Hi again,

Just 1 last question. For part (ii) couldn't I just sub z=2itan(pi/10) to show that it satisfies the equation, which proves that it's a root?

Thanks

yibbon · Jul 23, 2009

jm01 said:
Hi again,

Just 1 last question. For part (ii) couldn't I just sub z=2itan(pi/10) to show that it satisfies the equation, which proves that it's a root?

Thanks

jetblack does this, however by saying z=2iTan(pi/10) you assume a value for theta, hence can simply sub the theta value into the RHS and find when Cos(5@) = 0. I think doing the expansion would take too long but would show it's a root yes

cutemouse · Jul 23, 2009

Well, if I sub z=2itanpi/10, then I can just use the result from part (i), which yields zero. Nothing too hard about that lol.

yibbon · Jul 23, 2009

jm01 said:
Well, if I sub z=2itanpi/10, then I can just use the result from part (i), which yields zero. Nothing too hard about that lol.

Same thing

Hard complex number question (1 Viewer)

Account Closed

et tu

Administrator

Banned

Banned

Account Closed

Active Member

Active Member

Retired

Drover

Administrator

Banned

Banned

Active Member

Banned

Account Closed

Account Closed

Member

Account Closed

Member

Users Who Are Viewing This Thread (Users: 0, Guests: 1)