Basic Math Problem of the Week 11/09/2017

  • Context: Challenge 
  • Thread starter Thread starter PF PotW Robot
  • Start date Start date
Click For Summary

Discussion Overview

The discussion revolves around evaluating the expression $$\cos^3 \left(\dfrac{2\pi}{7}\right)+\cos^3 \left(\dfrac{4\pi}{7}\right)+\cos^3 \left(\dfrac{8\pi}{7}\right)$$ without using a calculator. Participants explore various methods and approaches to simplify the expression, sharing insights and techniques related to trigonometric identities and polynomial theory.

Discussion Character

  • Exploratory
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • One participant performed a numerical check and found a value of -0.5, suggesting an elegant simplification exists.
  • Another participant noted that the angles sum to a full circle and attempted to use the identity $$4\cos^3 \varphi= 3\cos \varphi + \cos 3 \varphi$$ to simplify the expression but felt they were missing a clue.
  • A different participant proposed using the exponential form of cosine to further manipulate the expression.
  • One participant introduced symmetry in their approach, defining variables for the cosines and applying identities to relate them, suggesting that the sum of the cosines equals the sum of their cubes.
  • Another participant referenced a video that provided additional identities related to the problem, noting a difference in the angles used compared to the video.
  • One participant detailed a method involving polynomial and matrix theory, discussing the properties of the companion matrix and symmetric sums to derive the solution.

Areas of Agreement / Disagreement

Participants express various approaches and insights, but there is no consensus on a single method or solution. Multiple competing views and techniques are presented, indicating an ongoing exploration of the problem.

Contextual Notes

Some participants mention identities and relationships that depend on specific assumptions about the angles and their properties, which may not be universally accepted or resolved within the discussion.

PF PotW Robot
Here is this week's basic math problem. We have several members who will check solutions, but we also welcome the community in general to step in. We also encourage finding different methods to the solution. If one has been found, see if there is another way. Using spoiler tags is optional. Occasionally there will be prizes for extraordinary or clever methods.

Without using a calculator, evaluate $$\cos^3 \left(\dfrac{2\pi}{7}\right)+\cos^3 \left(\dfrac{4\pi}{7}\right)+\cos^3 \left(\dfrac{8\pi}{7}\right)$$

(PotW thanks to our friends at http://www.mathhelpboards.com/)
 
  • Like
Likes   Reactions: Charles Link
Physics news on Phys.org
I did a numerical check of the expression and got a value of -0.5 so that means there is some elegant way of simplifying it.
 
  • Like
Likes   Reactions: QuantumQuest and Charles Link
And the angles sum up to a full circle. I tried ##4\cos^3 \varphi= 3\cos \varphi + \cos 3 \varphi## to get rid of the powers, but I'm missing the clue. The sum becomes with this formula
$$
\cos^3 \dfrac{2}{7}\pi + \cos^3 \dfrac{4}{7}\pi + \cos^3 \dfrac{8}{7}\pi = \cos \dfrac{2}{7}\pi + \cos \dfrac{4}{7}\pi + \cos \dfrac{8}{7}\pi
$$
I assume the rest can be done by ##\cos \varphi = \dfrac{1}{2} (\exp i\varphi + \exp -i\varphi)##.
 
Last edited:
  • Like
Likes   Reactions: QuantumQuest and Charles Link
fresh_42 said:
And the angles sum up to a full circle. I tried ##4\cos^3 \varphi= 3\cos \varphi + \cos 3 \varphi## to get rid of the powers, but I'm missing the clue. The sum becomes with this formula
$$
\cos^3 \dfrac{2}{7}\pi + \cos^3 \dfrac{4}{7}\pi + \cos^3 \dfrac{8}{7}\pi = \cos \dfrac{2}{7}\pi + \cos \dfrac{4}{7}\pi + \cos \dfrac{8}{7}\pi
$$
I assume the rest can be done by ##\cos \varphi = \dfrac{1}{2} (\exp i\varphi + \exp -i\varphi)##.
@fresh_42 That is very helpful. A google of cos(2 pi /7) showed a couple of identities with expressions like the one that you have that are solved by multiplying by ## \frac{\sin(\pi/7)}{\sin(\pi/7)} ##. See the video: ## \\ ## Additional comment:Our equation has an ## 8 \pi/7 ## instead of the ## 6 \pi/7 ## in the video. This makes the last terms in our case ## \sin(9 \pi /7)-\sin(7 \pi/7) ## . We have ## \sin(9 \pi /7)=-\sin(5 \pi/7) ##, so the results are identical.
 
Last edited:
  • Like
Likes   Reactions: fresh_42, QuantumQuest and jedishrfu
That’s a nice video And the solution is quite elegant.
 
  • Like
Likes   Reactions: Charles Link
Here is my approach:

let

##\lambda_1 = cos\big(\frac{2}{7}\pi\big)##
##\lambda_2 = cos\big(\frac{4}{7}\pi\big)##
##\lambda_3 = cos\big(\frac{8}{7}\pi\big)##

- - - -
key idea: symmetry

note ##cos(a)## obeys the identity: ##cos^2(a) = \frac{1}{2}\big(1+ cos(2a)\big)##

so if we apply this to each ##\lambda##, we get:
##\lambda_1^2 = \frac{1}{2}\big(1+ \lambda_2\big)##
##\lambda_2^2 = \frac{1}{2}\big(1+ \lambda_3\big)##
##\lambda_3^2 = \frac{1}{2}\big(1+ \lambda_1\big)##

Also noting that each ##\lambda_k## is the cosine of some multiple of pi, where each multiple is *not* a natural number, we know that ##\big \vert \lambda_k \big \vert \lt 1## for ##k \in \{1,2,3\}## -- i.e. each lambda has magnitude of less than one.

as mentioned by @fresh_42 we know that

##\lambda_1 + \lambda_2 + \lambda_3 = \lambda_1^3 + \lambda_2^3 + \lambda_3^3##

(which translates to the relation below)

##trace\big(\mathbf C\big) = trace\big(\mathbf C^3\big)##

- - - -
Solution:
From here I just beat this problem with a hammer using some facts from polynomial and matrix theory -- specifically things related to symmetric sums.

These ##\lambda##'s are all real numbers -- let them be interpreted as roots to some polynomial in monic form, i.e.

##p(x) = x^3 + a_2 x^2 + a_1 x + a_0 = (x- \lambda_1)(x- \lambda_2)(x-\lambda_3)##

We can encode this polynomial in the companion matrix ##\mathbf C##.

we know that

##a_2 = -trace\big(\mathbf C\big)##Newton's Identities (or just mechanical manipulation of the companion matrix) tell us that

##a_1 = -\frac{1}{2}\Big(a_2 trace\big(\mathbf C\big) + trace\big(\mathbf C^2\big)\Big) = -\frac{1}{2}\Big( -trace\big(\mathbf C\big)^2 + trace\big(\mathbf C^2\big)\Big) = \frac{1}{2} trace\big(\mathbf C\big)^2 - \frac{1}{2} trace\big(\mathbf C^2\big) ##

but we also know that ##a_1## is a symmetric function given by

##a_1 = \lambda_1\lambda_2 + \lambda_2\lambda_3 + \lambda_1\lambda_3##

Now the symmetry exploit creeps in:

##trace\big(\mathbf C^3\big) = \lambda_1^3 + \lambda_2^3 + \lambda_3^3 = \lambda_1(\lambda_1^2) + \lambda_2(\lambda_2^2) + \lambda_3(\lambda_3^2) = \lambda_1\big(\frac{1}{2}( 1 + \lambda_2)\big) + \lambda_2\big(\frac{1}{2}( 1 + \lambda_3)\big) + \lambda_3\big(\frac{1}{2}( 1 + \lambda_1)\big) ####trace\big(\mathbf C\big) = trace\big(\mathbf C^3\big) = \frac{1}{2}\big(\lambda_1\lambda_2 + \lambda_2\lambda_3 + \lambda_3 \lambda_1\big) +\frac{1}{2}\big(\lambda_1 + \lambda_2+ \lambda_3 \big) = \frac{1}{2}\big(\lambda_1\lambda_2 + \lambda_2\lambda_3 + \lambda_3 \lambda_1\big) +\frac{1}{2}trace\big(\mathbf C \big)##

##\frac{1}{2} trace\big(\mathbf C\big) = \frac{1}{2} \big(\lambda_1\lambda_2 + \lambda_2\lambda_3 + \lambda_3 \lambda_1\big)##

## trace\big(\mathbf C\big) = a_1 = trace\big(\mathbf C^3\big)##

using the Newton identity, this gives us:

##a_1 = trace\big(\mathbf C\big) = \frac{1}{2} trace\big(\mathbf C\big)^2 - \frac{1}{2} trace\big(\mathbf C^2\big) ##

- - - -

and we just need to make ##trace\big(\mathbf C^2\big)## 'go away'.

To finish this off, we examine the eigenvalue representation of ##trace\big(\mathbf C^2\big)## and again apply the symmetry exploit:

##trace\big(\mathbf C^2\big) = \lambda_1^2 + \lambda_2^2 + \lambda_3^2 = \frac{1}{2}\big(1 + \lambda_2\big) + \frac{1}{2}\big(1 + \lambda_3\big) + \frac{1}{2}\big(1 + \lambda_1\big) = \frac{3}{2} +\frac{1}{2}\big(trace\big(\mathbf C\big)\big)##

Now make the substitution into our equation from the Newton Identity.

##a_1 = trace\big(\mathbf C\big) = \frac{1}{2} trace\big(\mathbf C\big)^2 - \frac{1}{2} \Big(trace\big(\mathbf C^2\big)\Big) = \frac{1}{2} trace\big(\mathbf C\big)^2- \frac{1}{2}\Big( \frac{3}{2} +\frac{1}{2}trace\big(\mathbf C\big)\Big) ##

multiply everything by 2

##2 a_1 = a_1^2- \Big( \frac{3}{2} +\frac{1}{2}a_1 \Big) ##

##0 = a_1^2 - \frac{5}{2}a_1 - \frac{3}{2}##

And we can eyeball the trace and determinant of such an equation as having roots of ##-\frac{1}{2}## and ##3##. Because each ##\lambda## has magnitude less than one, by triangle inequality, we know

##\big \vert \lambda_1 + \lambda_2 + \lambda_3 \big \vert\leq \big \vert \lambda_1\big \vert + \big \vert\lambda_2\big \vert + \big \vert\lambda_3\big \vert \lt 3##

and hence we throw out the root of 3.

We keep the root of

##-\frac{1}{2} = a_1 = trace\big(\mathbf C\big) =trace\big(\mathbf C^3\big) = \lambda_1^3 + \lambda_2^3 + \lambda_3^3##

as our answer, which completes the proof.
 
Last edited:
  • Like
Likes   Reactions: Charles Link and jim mcnamara

Similar threads

  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 11 ·
Replies
11
Views
2K
  • · Replies 21 ·
Replies
21
Views
4K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K