[Paper Review] Method of Hidden Parameters and Pell's Equation

S. N. Arteha|ArXiv.org|Mar 14, 2003

Neural Networks and Applications8 references3 citations

TL;DR

This paper introduces the 'method of hidden parameters' to derive explicit parametric formulas for minimal solutions of Pell's equation $x^2 - Ay^2 = 1$, significantly reducing computation time by relating solutions across different $A$ values through fixed parameters $l$ and $m$. The method enables direct computation of infinite families of solutions using small auxiliary parameters, cutting calculation time by up to 1.83 times compared to standard continued fraction methods.

ABSTRACT

Using the representation of numbers, the methods of decreasing the number of calculation steps for Pell's equation are developed. The formulae relating the natural solution x0 of Pell's equation for A0 with the infinite number of minimal natural solutions xi for other Ai are obtained for various cases. The parametric representation of Pell's equation solutions are obtained with the help of new "method of hidden parameters".

Motivation & Objective

• To reduce the number of calculation steps required to solve Pell’s equation $x^2 - Ay^2 = 1$ for arbitrary non-square $A$.
• To derive explicit parametric formulas relating the minimal solution $x_0$ for a given $A_0$ to an infinite family of minimal solutions $x_i$ for other $A_i$.
• To identify and exploit structural patterns in solutions—particularly for prime and 'quasiprime' $A$ values—where solution sizes are maximized.
• To develop a systematic method, the 'method of hidden parameters', that transforms the solution process into a parameterized identity by introducing and eliminating auxiliary parameters.

Proposed method

• Introduces two types of relations: 'horizontal' relations, where basic parameters $l$ and $m$ are fixed and $x_i$ varies linearly with index $i$, and 'vertical' relations, where $l$ and $m$ vary nonlinearly.
• Uses parametric representations of solutions via $a_i = a_0 + i|l^2 - 3m^2|$, $b_i = b_0 + 2ilm$, and $A_i = a_i^2 + 3b_i^2$ to generate infinite solution families.
• Applies the identity $(y^2 - Ax^2)(Y^2 - AX^2) = (yY \pm AxX)^2 - A(yX \pm xY)^2$ to derive and verify parametric identities.
• Employs recursive sequences $B_n = d_n B_{n-1} + B_{n-2}$ with $B_{-2} = 1$, $B_{-1} = 0$ to link continued fraction convergents to solution parameters.
• Uses the condition $dl - 3gm = \pm 1$ to parameterize initial values $a_0$ and $b_0$ in terms of $d$, $g$, $l$, and $m$ for constructing solution families.
• Reduces the original problem by introducing hidden parameters, deriving relations among them, and eliminating them to express solutions in terms of initial $A_0$ and $x_0$.

Experimental results

Research questions

• RQ1 What parametric structure underlies the minimal solutions of Pell’s equation across different $A$ values, especially for prime and 'quasiprime' $A$?
• RQ2 How can the number of computational steps in solving Pell’s equation be reduced by exploiting structural patterns in solutions?
• RQ3 Can infinite families of solutions for Pell’s equation be generated from a single known solution via explicit parametric formulas?
• RQ4 What is the role of the indeterminate linear equation $ax - by = 1$ in connecting Pell’s equation to the method of hidden parameters?
• RQ5 How do the local and absolute maxima of solution sizes $x_i$ correlate with the arithmetic structure of $A_i$?

Key findings

• The method of hidden parameters enables explicit parametric representation of Pell’s equation solutions as identities, linking a single solution $x_0$ for $A_0$ to infinitely many solutions $x_i$ for other $A_i$.
• For $A < 1,000,000$, the modified algorithm using the method reduced computation time from 3h 2m (standard continued fraction) to 1h 43m, a speedup of 1.827x.
• For $A < 2,000,000$, the speedup increased to 1.828x, demonstrating that the method is particularly effective for the most difficult cases (e.g., absolute maxima).
• All absolute maxima for $A > 46$ occur at $A = 4n+1$ primes, and local maxima occur at primes $p_i$ or double primes $2p_j$, suggesting structural patterns in solution complexity.
• The method identifies 'quasiprime' $A$ values (e.g., $A = 1729$) that behave like primes in solution structure but lack the standard form in the sequential differences method.
• The parametric solution for the case $x = 2(l^2 + 3m^2)$ is valid when $A = A_0^2 + 3B_0^2 = 4N+1$ with $B_0$ even, and uses substitutions $a = A_0 \pm 3B_0$, $b = A_0 \mp B_0$ to generate solutions.