I wonder whether there is a pattern that goes on and on: $$(a+b)\,(a-b) = a^2 - b^2$$ $$(a+b)\,(a+(-1/2 + i \sqrt{3}/2)b)\,(a+(-1/2 - i \sqrt{3}/2)b) = a^3 + b^3$$ $$(a+b)\,(a+i b)\,(a-b)\,(a-i b) = a^4 - b^4$$ The general product would be as follows where $\epsilon = e^{2 i \pi/n}$ is the n-th unit root: $$\prod_{k=0}^{n-1}(a+\epsilon^k b) =\,?$$

  • 2
    $\begingroup$ $a^n-(-b)^n\phantom{}$. $\endgroup$ – Lord Shark the Unknown Jun 23 '17 at 20:26
  • 1
    $\begingroup$ Any proof for that? It looks like, but I am not sure. $\endgroup$ – j4n bur53 Jun 23 '17 at 20:28
  • 1
    $\begingroup$ Use \prod for $\prod$. It works better than \Pi :) $\endgroup$ – Cameron Williams Jun 23 '17 at 20:37
  • $\begingroup$ Hint: $\prod_{k=0}^{n-1}(a+\epsilon^k b) = (-1)^nb^n \prod_{k=0}^{n-1}\left(\frac{a}{b}-\epsilon^k\right)=(-1)^nb^nP\left(\frac{a}{b}\right)$ where $P(z)$ is the polynomial with roots $1, \epsilon, \epsilon^2, \cdots, \epsilon^{n-1}\,$. $\endgroup$ – dxiv Jun 23 '17 at 20:37

Your problem is equivalent to find the roots of


consider $b>0$. So, the roots are $$-b\cdot (\text {roots of unit})$$

and once you can split $p(a)$ as:


where $a_i$ is a root of $p(a)$ then you get what you want.


Your Answer

By clicking "Post Your Answer", you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.