The sets A = { z : z 1 8 = 1 } and B = { w : w 4 8 = 1 } are both sets of complex roots of unity. The set C = { z w : z ∈ A , w ∈ B } is also a set of complex roots of unity. How many distinct elements are in C ?
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.
C is just the set of 1 4 4 th roots of unity. So the answer is 1 4 4 .
Clearly any element of C is a 1 4 4 th root of unity. Now let a = e x p ( 2 π i / 1 4 4 ) , so that any 1 4 4 th root of unity is a n for some n . Write n = 8 x + 3 y for some integers x , y ; this is always possible because 8 and 3 are relatively prime. Then a n = ( a 8 ) x ( a 3 ) y shows that a n is a product of an 1 8 th root of unity and a 4 8 th root of unity, so a n ∈ C . So all the 1 4 4 th roots of unity are in C .
The same argument shows that, if 1 8 and 4 8 are replaced by m and n , the answer is lcm ( m , n ) .
Hi , can you tell me how did you get 144 ?
Log in to reply
It's the LCM of 1 8 and 4 8 .
Log in to reply
And why did you take n = 8 x + 3 y ?
Hi can you help me out ?
can you expand on the solution? it looks great, but i'm too dumb, i need some clarification. thanks!!!
Problem Loading...
Note Loading...
Set Loading...
Consider the general case: for sets A n = { z : z n = 1 } and B m = { w : w m = 1 } , how many elements are there in C n , m = { z w : z ∈ A n , w ∈ B m } ?
Since the sets A n and B m are sets of roots of unity we can rewrite them like this: A n B m = { e n 2 π i : i = 1 , 2 , … , n } = { e m 2 π j : j = 1 , 2 , … , m }
Now, this means we can rewrite C n , m as: C n , m = { e 2 π ( n i + m j ) : i = 1 , 2 , … , n ; j = 1 , 2 , … , m }
That is, every element x ∈ C n , m can be expressed as: x = e 2 π m n i m + j n for some i , j . Clearly, then, every element of C is a mn-th root of unity. Let g = g cd ( m , n ) , and N = n / g , M = g / m , then this can be rewritten as: x = e 2 π M N i M + j N That is, C contains only MN-th roots of unity. The question now is whether it contains all such roots. That is, what values can i M + j N take?
Since M and N are coprime, Bézout's identity tells us that x M + y N = k has integer solutions for x,y for any k. Let i k ′ , j k ′ be one such solution, for each value of k ∈ { 1 , 2 , … , M N } . That is, i k ′ M + j k ′ N = k Let i = i k ′ ( m o d M N ) and j = j k ′ ( m o d M N ) , then i M + j N = k + r M N for some r ∈ Z . Returning to our expression for x , this then becomes: x = e 2 π M N k + r M N = e 2 π ( M N k + r ) = e 2 π M N k
That is, if α = e 2 π / M N , β = e 2 π / n , and γ = e 2 π / m then for any value k ∈ { 1 , 2 , … , M N } there is a suitable value of i ∈ { 1 , 2 , … , n } and j ∈ { 1 , 2 , … , m } such that α k = β i γ j . And since β ∈ A n and γ ∈ B m , this shows that every possible MN-th root of unity must be in C n , m .
There are MN such roots, and so \DeclareMathOperator \lcm l c m ∣ C n , m ∣ = M N = \lcm ( n , m ) .
In this problem, n = 1 8 , m = 4 8 so \DeclareMathOperator \lcm l c m ∣ C 1 8 , 4 8 ∣ = \lcm ( 1 8 , 4 8 ) = 1 4 4 .