0
$\begingroup$

A Pell’s equation is a diophantine equation in $x$ and $y$ of the form $$x^2-d\cdot y^2=1$$ with $d$ a square-free integer.

The fundamental solution of a Pell’s equation is the smallest (with the smallest absolute value) number $$z_1=x_1+y_1\cdot\sqrt{d}$$ for which $(x,y)$ is a solution to the Pell’s equation and $|z_1|>1$.

In this case, I know (and understand) that the general solution to the equation is $z_n=z_1^n$.


The general form of Pell’s equation is a diophantine equation of the form $$r\cdot x^2 - d\cdot y^2=c$$ for which $r>0$, $d>0$ and $c\neq0$ are integers.

And now, my book says that a (there can be multiple) fundamental solution to this generalised form is a number $$z_0=x_0+y_0\sqrt{d}$$ for which $x_0^2-d\cdot y_0^2=c$ and $1<|x_0+y_0\sqrt{d}|<z_1$. Here I assume that $z_1$ is the fundamental solution of the corresponding ‘simple’ Pell’s equation $x^2-d\cdot y^2=1$.

Next, it should be possible to calculate the general solution $z_n=z_0\cdot z_1^n$.


However, I don’t understand how to calculate a fundamental solution of the generalised equation and why the general term should equal $z_0\cdot z_1^n$.

For example, consider the diophantine equation $x^2-2y^2=-1$.

Then $z_1=3+2\sqrt{2}$, as $(3,2)$ is the fundamental solution of $x^2-2y^2=1$.

Now, I would expect that there are no fundamental solutions to the original equation, as there are no numbers $z_0=x_0+y_0\sqrt{d}$ for which $(x_0,y_0)$ is a solution to $x^2-2y^2=-1$ and $1<|z_0|<|z_1|$.

However, the original equation does have integer solutions! (take, for example, $(x,y)=(7,5)$

Could someone please explain me what I’m doing wrong?

$\endgroup$
  • $\begingroup$ What are $z_0,z_1$ in inequality (after "and" of third last sentence) and why is that inequality imposed? $\endgroup$ – coffeemath Apr 2 at 19:25
  • $\begingroup$ what book are you using? $\endgroup$ – Will Jagy Apr 2 at 20:12
2
$\begingroup$

In your example, $1-\sqrt{2}$ generates the solution $x=1,y=-1$ and this does satisy the fundamental solution inequality. The equivalent $z_0 = \sqrt{2}-1$ is easier to work with. With $z_1 = 3+2\sqrt{2}$, this $z_0$ generates the following solutions: $$ z_0z_1 = 1+\sqrt{2} \Longrightarrow (x=1,y=1) \\ z_0z_1^2 = 7+5\sqrt{2} \Longrightarrow (x=7,y=5) \\ z_0z_1^3 = 41+29\sqrt{2} \Longrightarrow (x=41,y=29) \\ z_0z_1^3 = 239+169\sqrt{2} \Longrightarrow (x=239,y=169) \\ \vdots $$ Nothing cute tells me how to find that first $z_0$ in general, but that is the same situation as for the original Pell equation, where nothing tells you how to find the smallest $z_1$. In fact, the situation is better here because at least you know a limit on how large $z_0$ can be.

I don't know that an exhaustive search revealing no $z_0$ in the range $1 \leq |x_0+y_0\sqrt{d}| < z_1$ proves that the corresponding generalized Pell equation has no solution, but I suspect that is the case.

$\endgroup$
  • $\begingroup$ I think that z_0 should be 1+\sqrt{2}, because |z_0| must be larger than 1. But (at least in this example) it doesn’t matter, as it generates the same solutions as z_0 = \sqrt{2}-1. (This is because (\sqrt{2}-1)(3+2\sqrt{2}) = 1+\sqrt{2}.) Still, thanks for the answer, it really cleared things up a bit. $\endgroup$ – Jonas De Schouwer Apr 2 at 21:15

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.