I came across this observation while trying to answer this post using the Pell equation $x^2-2y^2=1$. Define,

$$P(m) = \frac{ (3+2\sqrt{2})^m+(3-2\sqrt{2})^m}{2}$$

$$Q(m) = \frac{ (3+2\sqrt{2})^m-(3-2\sqrt{2})^m}{2\sqrt{2}}$$

such that $P(m)^2-2Q(m)^2=1$. We have,

$$P(2^k) = \color{brown}{3, 17, 577, 665857, 886731088897,\dots}$$

$$Q(2^k) = \color{blue}{2, 12, 408, 470832, 627013566048,\dots}$$

Notice that,

$$\color{blue}{12} = 2^2\cdot\color{brown}3$$

$$\color{blue}{408} = 2^3\cdot\color{brown}{3\cdot17}$$

$$\color{blue}{470832} = 2^4\color{brown}{\cdot3\cdot17\cdot577}$$

and so on. In general, given the roots of $ax^2+bx+c=0$,

$$x_{1,2} = \frac{-b\pm\sqrt{b^2-4ac}}{2a}$$

and let $D=b^2-4ac$.

Question: Is it true that,

$$\frac{2^{n}}{a}\prod_{k=0}^n \frac{x_1^{2^k}+x_2^{2^k}}{2}=\frac{x_1^{2^{n+1}}-x_2^{2^{n+1}}}{2\sqrt{D}}$$


Note that $x^2-y^2=(x-y)(x+y)$ and by induction $$x^{2^{n+1}}-y^{2^{n+1}}=(x^{2^n}-y^{2^n})(x^{2^n}+y^{2^n})=(x-y)\prod_{k=0}^n(x^{2^k}+y^{2^k}).$$ For roots $x_1, x_2$ of a quadratic equation and $x_1\leq x_2$ we have $x_2-x_1 = \sqrt{D}/a$. This proves your identity.

| cite | improve this answer | |
  • $\begingroup$ I didn't expect the reason would be so mundane. :) $\endgroup$ – Tito Piezas III Dec 25 '15 at 8:29

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.