Dot Products Around the Octagon

Geometry Level 3

Consider the regular octagon centered at the origin as shown above. Eight unit vectors are drawn from the center of the octagon to each of its vertices. For each pair of distinct unit vectors, the dot product is computed. What is the sum of all of these dot products?


The answer is -4.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

4 solutions

Andrew Ellinor
Oct 2, 2015

There are a total of ( 8 2 ) 8\choose2 = 28 = 28 pairs of unit vectors: 8 pairs make an angle of 4 5 45^\circ , 8 pairs make an angle of 9 0 90^\circ , 8 pairs make an angle of 13 5 135^\circ , and 4 pairs make an angle of 18 0 180^\circ . Using the formula u v = u v cos ( θ ) \vec{u}\cdot\vec{v} = \|u\|\|v\|\cos(\theta) and since each vector is has magnitude 1, the contribution by each type is:

8 cos ( 4 5 ) = 4 2 8 cos ( 9 0 ) = 0 8 cos ( 13 5 ) = 4 2 4 cos ( 18 0 ) = 4 \begin{aligned} 8\cos(45^\circ) &= 4\sqrt{2} \\ 8\cos(90^\circ) &= 0 \\ 8\cos(135^\circ) &= -4\sqrt{2} \\ 4\cos(180^\circ) &= -4 \end{aligned}

The sum of all these values is 4 \boxed{-4} .

It is superb solution. Thank you.

Dev Rajyaguru - 5 years, 8 months ago

A (B + C) = A B + A C is not actually an ever true identity though. When involving with an upright in general number (3 dimensional number); it has to be true only occasionally, when B and C is said to be co-directional, where their complex portion is of exactly sole directional the same, not even relation of -1 or exact opposite is allowed. Otherwise, the logic is incorrect instead. This also explains why Complex functions such as trigonometric and hyperbolic plus their inverses are invalid to 3 dimensional numbers as their derivatives based on this identity ought to fail. Only Ln G is one step definable as | G | + j Q + k P.

Lu Chee Ket - 5 years, 8 months ago
Calvin Lin Staff
Oct 2, 2015

Let the vectors be represented by v i v_i .

We have v i = 0 \sum v_i = 0 (since pairs of them cancel out).
We also have v i 2 = 8 \sum v_i ^2 = 8 since each of them are unit vectors.
Since

( v i ) 2 = v i 2 + 2 i < j v i v j . ( \sum v_i ) ^2 = \sum v_i^2 + 2 \sum_{i < j} v_i v_j.

Hence, we conclude that v i v j = 0 8 2 = 4 \sum v_i v_j = \frac{0-8}{2} = -4 .


Note: All that we needed was v i = 0 \sum v_i = 0 . We didn't need to know the exact shape of the octagon.

Could u explain where the 3rd equation comes from?? Thanks!

Ivan Martinez - 5 years, 8 months ago

Log in to reply

(Made slight edits.)

What happens when you square v i \sum v_i ? You get the square of the individual terms which are v i 2 \sum v_i ^2 , and you also get the cross terms which are v i v j \sum v_i v_j with j i j \neq i .

Note that we only want each pair once, and we don't want both v i v j v_i v_j and v j v i v_j v_i to appear in the summation. Hence this accounts for the coefficient of 2.

Calvin Lin Staff - 5 years, 8 months ago

Log in to reply

Oh, I get it now, thanks a lot!

Ivan Martinez - 5 years, 8 months ago

This is powerful provided the equation is true.

Lu Chee Ket - 5 years, 8 months ago

Let the vectors be v 1 , , v 8 v_1, \dots, v_8 . For any given vector v i v_i we have j v i v j = v i ( j v j ) = v i 0 = 0. \sum_j v_i\cdot v_j = v_i\cdot\left(\sum_j v_j\right) = v_i\cdot 0 = 0. But we must sum pairs of distinct vectors, so we subtract the product of v i v_i with itself: j i v i v j = ( j v i v j ) v i v i = 0 v i 2 = 1. \sum_{j\neq i} v_i\cdot v_j = \left(\sum_j v_i\cdot v_j\right) - v_i\cdot v_i = 0 - v_i^2 = -1. Adding this for all eight vectors we get i j i v i v j = 8 ( 1 ) = 8. \sum_i \sum_{j\neq i} v_i\cdot v_j = 8 (-1) = -8. However, now we have counted each pair i j i\neq j twice, so divide by two. The answer is 8 2 = 4 \frac{-8}2 = -4 .

Fun solution!

Andrew Ellinor - 5 years, 8 months ago
Stewart Gordon
Oct 3, 2015

The dot product of vectors is distributive over addition. As such, for a given pair ( v i , v j ) (v_i, v_j) , the two products v i . v j v_i . v_j and v i . v j v_i . -v_j cancel out, except that we have an unpaired term where v i = v j v_i = -v_j (since only distinct pairs are counted).

So it's just v 1 . v 5 + v 2 . v 6 + v 3 . v 7 + v 4 . v 8 v_1 . v_5 + v_2 . v_6 + v_3 . v_7 + v_4 . v_8 . Each of these dot products is equal to 1 × 1 cos π = 1 1 \times 1 \cos \pi = -1 , giving a total of 4 \boxed{-4} .

A (B + C) = A B + A C is not actually an ever true identity though. When involving with an upright in general number (3 dimensional number); it has to be true only occasionally, when B and C is said to be co-directional, where their complex portion is of exactly sole directional the same, not even relation of -1 or exact opposite is allowed. Otherwise, the logic is incorrect instead. This also explains why Complex functions such as trigonometric and hyperbolic plus their inverses are invalid to 3 dimensional numbers as their derivatives based on this identity ought to fail. Only Ln G is one step definable as | G | + j Q + k P.

Lu Chee Ket - 5 years, 8 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...