I am struggling to understand how you use continous fractions to solve it and why D is relevant. I know how to write P/Q in continous fraction notation but I am not sure how to express irrational numbers.
I am struggling to understand how you use continous fractions to solve it and why D is relevant. I know how to write P/Q in continous fraction notation but I am not sure how to express irrational numbers.
The reason that D is important is that you can re-cast the problem as finding solutions to
(x+yD)(x−yD)=1
so that the problem is equivalent to finding a non-trivial unit of norm 1 in the ring Z[D].
It sounds as though you need to brush up on the theory of continued fractions. There's a pretty good tutorial here - and it includes a treatment of the key fact about how the continued fraction of a quadratic surd (like D) is repeating.
The reason that D is important is that you can re-cast the problem as finding solutions to
(x+yD)(x−yD)=1
so that the problem is equivalent to finding a non-trivial unit of norm 1 in the ring Z[D].
It sounds as though you need to brush up on the theory of continued fractions. There's a pretty good tutorial here - and it includes a treatment of the key fact about how the continued fraction of a quadratic surd (like D) is repeating.
Thanks, that has cleared some things up. This is something I haven't learnt yet, it was a problem for the york applicant day. Could you explain what a non trivial unit of norm 1 in the ring z is?
Thanks, that has cleared some things up. This is something I haven't learnt yet, it was a problem for the york applicant day. Could you explain what a non trivial unit of norm 1 in the ring z is?
OK. So
Z[D]={x+yD:x,y∈Z}
with the obvious operations of addition and multiplication to make it a ring. The point is that we use numbers of the form x+yD to study solutions of the Pell equation.
The norm of x+yD is defined to be precisely x2−Dy2 (if you think about the norm of a complex number being derived from the product of the number with its complex conjugate, you get an analogy of where this is coming from).
So if the norm of x+yD is x2−Dy2 then to solve the Pell equation is precisely to find a conjugate pair of elements satisfying
(x+yD)(x−yD)=1
In other words, we want to find divisors of unity (so-called units) that are not equal to unity (i.e. they are non-trivial) and which have norm one.
Are you thinking of coming to York? Good department - but I'm biased as I work at York Uni!
I really liked it there, the research from the maths department really impressed me. It is definitely going to be a university I choose as an option, just not sure whether insurance or firm.
with the obvious operations of addition and multiplication to make it a ring. The point is that we use numbers of the form x+yD to study solutions of the Pell equation.
The norm of x+yD is defined to be precisely x2−Dy2 (if you think about the norm of a complex number being derived from the product of the number with its complex conjugate, you get an analogy of where this is coming from).
So if the norm of x+yD is x2−Dy2 then to solve the Pell equation is precisely to find a conjugate pair of elements satisfying
(x+yD)(x−yD)=1
In other words, we want to find divisors of unity (so-called units) that are not equal to unity (i.e. they are non-trivial) and which have norm one.
I'm not entirely sure what you are doing here, but it looks as if you are calculating the general continued fraction rather than the simple form we need here. If you are following the link I posted earlier, you should be working through the algorithm in section 6.2.2.
So you've lost the last term of the recurrence. After this, the terms of the continued fraction just repeat.
From here the point is that you get the first solution of the Pell equation from among the convergents of the continued fraction representation of 41.
x2−ny2=1
if and only if yx is among the convergents of the continued fraction representation of n. So you just have to start churning through the convergents of the continued fraction representation until you find a solution. Here, you should get convergents starting 6, 13/2, 32/5...
If you want to carry through with the algebra, it might help to refer to here for some computational help.