Respuesta :
Let [tex]n[/tex] be a fixed positive integer. Then [tex]n[/tex] points can be equally spaced along the boundary of the unit disk in the complex plane, and their positions described in the complex plane are given by [tex]e^{i2\pi k/n}[/tex], where [tex]k=0,1,\ldots,n-1[/tex].
So, for any [tex]0\le k\le n-1[/tex], we can write
[tex]A_k=e^{i2\pi k/n}=\cos\dfrac{2\pi k}n+i\sin\dfrac{2\pi k}n\equiv\left(\cos\dfrac{2\pi k}n,\sin\dfrac{2\pi k}n\right)[/tex]
The distance between two points [tex]A_j[/tex] and [tex]A_k[/tex] is given by the norm of the difference of the points. We denote this by
[tex]A_jA_k:=|A_j-A_k|[/tex]
We're interested in all the possible distinct pairings, which means [tex]A_jA_k[/tex] and [tex]A_kA_j[/tex] are to be considered the same, so to accomplish this we assume [tex]0\le j<k\le n-1[/tex].
Then the sum we're interested in finding is
[tex]\displaystyle\sum_{0\le j<k\le n-1}(A_jA_k)^2=(A_0A_1)^2+\cdots+(A_0A_{n-1})^2+(A_1A_2)^2+\cdots+(A_1A_{n-1})^2+(A_2A_3)^2+\cdots+(A_{n-2}A_{n-1})^2[/tex]
where this sum is identical to the one in your original question, except the indices of each point are shifted down by one.
We have
[tex](A_jA_k)^2=|A_j-A_k|^2[/tex]
[tex]=\left|\left(\cos\dfrac{2\pi j}n,\sin\dfrac{2\pi j}n\right)-\left(\cos\dfrac{2\pi k}n,\sin\dfrac{2\pi k}n\right)\right|^2[/tex]
[tex]=\left(\cos\dfrac{2\pi j}n-\cos\dfrac{2\pi k}n\right)^2+\left(\sin\dfrac{2\pi j}n-\sin\dfrac{2\pi k}n\right)^2[/tex]
[tex]=\cos^2\dfrac{2\pi j}n-2\cos\dfrac{2\pi j}n\cos\dfrac{2\pi k}n+\cos^2\dfrac{2\pi k}n+\sin^2\dfrac{2\pi j}n-2\sin\dfrac{2\pi j}n\sin\dfrac{2\pi k}n+\sin^2\dfrac{2\pi k}n[/tex]
[tex]=2-2\cos\dfrac{2\pi}n(j-k)[/tex]
[tex]=4\sin^2\dfrac\pi n(j-k)[/tex]
So we can express the sum as
[tex]\displaystyle\sum_{0\le j<k\le n-1}4\sin^2\dfrac\pi n(j-k)=\sum_{0\le j<k\le n-1}4\sin^2\dfrac\pi n(k-j)=\sum_{k=1}^{n-1}\sum_{j=0}^{k-1}4\sin^2\dfrac\pi n(k-j)[/tex]
For example, if we take [tex]n=4[/tex] points, we can see the points make up the vertices of a square with inradius 1, which means the distance between adjacent points ([tex]k=j+1[/tex]) is [tex]\sqrt2[/tex], while the distance between opposite points ([tex]k=j+2[/tex], which gets counted only once) is [tex]2[/tex]. So we would expect the sum to be
[tex](A_0A_1)^2+(A_0A_2)^2+(A_0A_3)^2+(A_1A_2)^2+(A_1A_3)^2+(A_2A_3)^2=4(\sqrt2)^2+2(2)^2=16[/tex]
In terms of our sum, we would add
[tex]\displaystyle\sum_{0\le j<k\le3}4\sin^2\dfrac\pi4(k-j)[/tex]
[tex]=4\left(4\sin^2\dfrac\pi4\right)+2\left(4\sin^2\dfrac{2\pi}4\right)[/tex]
[tex]=4(\text{distance between adjacent vertices})+2(\text{distance between opposite vertices})[/tex]
[tex]=16[/tex]
as required.
This was just an example to convince you that the formula works.
For [tex]n=12[/tex], you would end up with
[tex]4\sin^2\dfrac\pi{12}(1-0)+4\sin^2\dfrac\pi{12}(2-0)+\cdots+4\sin^2\dfrac\pi{12}(11-0)[/tex]
[tex]+4\sin^2\dfrac\pi{12}(2-1)+4\sin^2\dfrac\pi{12}(3-1)+\cdots+4\sin^2\dfrac\pi{12}(11-1)[/tex]
[tex]+4\sin^2\dfrac\pi{12}(3-2)+4\sin^2\dfrac\pi{12}(4-2)+\cdots+4\sin^2\dfrac\pi{12}(11-2)[/tex]
[tex]+\cdots[/tex]
[tex]+4\sin^2\dfrac\pi{12}(11-10)[/tex]
[tex]=144[/tex]
So, for any [tex]0\le k\le n-1[/tex], we can write
[tex]A_k=e^{i2\pi k/n}=\cos\dfrac{2\pi k}n+i\sin\dfrac{2\pi k}n\equiv\left(\cos\dfrac{2\pi k}n,\sin\dfrac{2\pi k}n\right)[/tex]
The distance between two points [tex]A_j[/tex] and [tex]A_k[/tex] is given by the norm of the difference of the points. We denote this by
[tex]A_jA_k:=|A_j-A_k|[/tex]
We're interested in all the possible distinct pairings, which means [tex]A_jA_k[/tex] and [tex]A_kA_j[/tex] are to be considered the same, so to accomplish this we assume [tex]0\le j<k\le n-1[/tex].
Then the sum we're interested in finding is
[tex]\displaystyle\sum_{0\le j<k\le n-1}(A_jA_k)^2=(A_0A_1)^2+\cdots+(A_0A_{n-1})^2+(A_1A_2)^2+\cdots+(A_1A_{n-1})^2+(A_2A_3)^2+\cdots+(A_{n-2}A_{n-1})^2[/tex]
where this sum is identical to the one in your original question, except the indices of each point are shifted down by one.
We have
[tex](A_jA_k)^2=|A_j-A_k|^2[/tex]
[tex]=\left|\left(\cos\dfrac{2\pi j}n,\sin\dfrac{2\pi j}n\right)-\left(\cos\dfrac{2\pi k}n,\sin\dfrac{2\pi k}n\right)\right|^2[/tex]
[tex]=\left(\cos\dfrac{2\pi j}n-\cos\dfrac{2\pi k}n\right)^2+\left(\sin\dfrac{2\pi j}n-\sin\dfrac{2\pi k}n\right)^2[/tex]
[tex]=\cos^2\dfrac{2\pi j}n-2\cos\dfrac{2\pi j}n\cos\dfrac{2\pi k}n+\cos^2\dfrac{2\pi k}n+\sin^2\dfrac{2\pi j}n-2\sin\dfrac{2\pi j}n\sin\dfrac{2\pi k}n+\sin^2\dfrac{2\pi k}n[/tex]
[tex]=2-2\cos\dfrac{2\pi}n(j-k)[/tex]
[tex]=4\sin^2\dfrac\pi n(j-k)[/tex]
So we can express the sum as
[tex]\displaystyle\sum_{0\le j<k\le n-1}4\sin^2\dfrac\pi n(j-k)=\sum_{0\le j<k\le n-1}4\sin^2\dfrac\pi n(k-j)=\sum_{k=1}^{n-1}\sum_{j=0}^{k-1}4\sin^2\dfrac\pi n(k-j)[/tex]
For example, if we take [tex]n=4[/tex] points, we can see the points make up the vertices of a square with inradius 1, which means the distance between adjacent points ([tex]k=j+1[/tex]) is [tex]\sqrt2[/tex], while the distance between opposite points ([tex]k=j+2[/tex], which gets counted only once) is [tex]2[/tex]. So we would expect the sum to be
[tex](A_0A_1)^2+(A_0A_2)^2+(A_0A_3)^2+(A_1A_2)^2+(A_1A_3)^2+(A_2A_3)^2=4(\sqrt2)^2+2(2)^2=16[/tex]
In terms of our sum, we would add
[tex]\displaystyle\sum_{0\le j<k\le3}4\sin^2\dfrac\pi4(k-j)[/tex]
[tex]=4\left(4\sin^2\dfrac\pi4\right)+2\left(4\sin^2\dfrac{2\pi}4\right)[/tex]
[tex]=4(\text{distance between adjacent vertices})+2(\text{distance between opposite vertices})[/tex]
[tex]=16[/tex]
as required.
This was just an example to convince you that the formula works.
For [tex]n=12[/tex], you would end up with
[tex]4\sin^2\dfrac\pi{12}(1-0)+4\sin^2\dfrac\pi{12}(2-0)+\cdots+4\sin^2\dfrac\pi{12}(11-0)[/tex]
[tex]+4\sin^2\dfrac\pi{12}(2-1)+4\sin^2\dfrac\pi{12}(3-1)+\cdots+4\sin^2\dfrac\pi{12}(11-1)[/tex]
[tex]+4\sin^2\dfrac\pi{12}(3-2)+4\sin^2\dfrac\pi{12}(4-2)+\cdots+4\sin^2\dfrac\pi{12}(11-2)[/tex]
[tex]+\cdots[/tex]
[tex]+4\sin^2\dfrac\pi{12}(11-10)[/tex]
[tex]=144[/tex]