# npm
> npm i polynomial-arithmetic
# pnpm
> pnpm add polynomial-arithmetic
# yarn
> yarn add polynomial-arithmetic
Library for performing polynomial arithmetic operations, designed in the context of polynomials in the Galois Field (GF). Test for primitiveness, irreducibility and setwise coprime taps, find the greatest common divisor polynomial besides multiplication, addition, subtraction, division and derivative.
Main use cases: Maximum length sequence (m-sequence) polynomial taps in Linear Feedback Shift Registers (LFSRs) (Cryptography, CRC, Digital Signal Processing, Testing and Benchmarking). Test primtive, irreducible, setwise coprime, use: extended euclidean algorithm for polynomials.
import { Polynomial } from "polynomial-arithmetic"
// Addition
const addend = new Polynomial("8x^4+32x-12");
const sum = addend.add("-x^5-16x+6")
console.log(sum.polyString) // - x^5 + 8x^4 + 16x - 6
// Multiplication
const term = new Polynomial("- x^5 + 8x^4 + 16x - 6")
const product = term.multiply("4x^2+8")
console.log(product.polyString) // - 4x^7 + 32x^6 - 8x^5 + 64x^4 + 64x^3 - 24x^2 + 128x - 48
// Subtraction
const minuend = new Polynomial("- 4x^7 + 32x^6 - 8x^5 + 64x^4 + 64x^3 - 24x^2 + 128x - 48")
const difference = minuend.sub("-4x^7 - 8x^5 + 64x^3 + 128x - 48")
console.log(difference.polyString) // 32x^6 + 64x^4 - 24x^2
// Division
const dividend = new Polynomial("32x^6 + 64x^4 - 24x^2")
const {quotient, remainder} = dividend.divide("8x^2")
console.log(quotient.polyString) // 4x^4 + 8x^2 - 3
// Greatest Common Divisor
const poly1 = new Polynomial("8x^4+32x-12")
const gcd = poly1.gcd(quotient) // - 3
console.log(gcd.polyString)
// Derivative
const poly = new Polynomial("8x^4+32x-12")
console.log(poly.derivative().polyString) // // 32x^3 + 32Long division of polynomials, enables computing modular arithmetic of a polynomial f(x) with a polynomial g(x) as modulus by dividing f(x) into g(x) and keeping the remainder. Furthermore if p(x) divides f(x) then the greatest common divisor is p(x) itself.
This long division method is optimized by computing the modulus (number) of intermediate results otherwise computed at the end, thus preventing the storage of large integer elements in the polynomial array, which may lead to incorrect results due to precision loss or NaN values.
import { FieldPolynomial } from "polynomial-arithmetic"
// Division
const dividend = new FieldPolynomial('x^9 + x^8 + x^7 + x^5 + x^4 + x^1 + 1');
const {quotient, remainder} = dividend.divide('x^4 + x^1 + 1');
console.log("div: ",quotient.polyString) // x^5 + x^4 + x^3 + x^2 + x
console.log("div: ",remainder.polyString) // 1
// XOR Addition
const addend = new FieldPolynomial('x^4 + x^3 + x^2 + x + 1');
const sum = addend.addGF(new FieldPolynomial('x^3 + x + 1'));
console.log(sum.polyString) // x^4 + x^2
// AND Multiplication
const term = new FieldPolynomial('x^4 + x^3 + x^2 + x + 1');
const product = term.multiplyGF('x^3 + x + 1');
console.log(product.polyString) // x^7 + x^6 + x^4 + x^3 + 1
// XOR Subtraction
const minuend = new FieldPolynomial("x^5 + x^3 + x^2 + 1");
const difference = minuend.subGF("x^2 + 1");
console.log(difference.polyString) // x^5 + x^3Bitwise optimized methods for polynomial arithmetic in GF(2) where coefficients are either 0 or 1, example: x^10 + x.Enabling efficient computation. Multiplication, bitwise AND, bit Carry-Less Product method. Subtraction & Addition both share the same method & result. XOR (bit exclusive OR) operation.
Long Division, Improved. Assuming polynomial in GF(2) thus working modulo 2 and removing the need of working with long arrays of coefficients. Instead working with short arrays of unique exponents, however, despite short lengths, elements in the array may reach large integers. Loops require less computation.
note: 2^53 - 1 highest integer before loosing precision
import { FieldPolynomial } from "polynomial-arithmetic"
const polynomial = new FieldPolynomial('x^4 + x^3 + x^2 + x + 1')
console.log(polynomial.isIrreducible()) // trueAn Irreducible polynomial over a given finite field (GF) (or domain) is a non-constant polynomial that cannot be factored into the product of two non-constant polynomials with coefficients in the same domain. Rabin's test of irreducibility, states (Fresham's Law of Exponents), n be the maximum possible period in F(pn), which is the size of the field, such that:
- xp^n - x divides f(x) mod (p)
- GCD(f(x),xp^(n/q) - x) mod(2) = 1 for each prime divisor q of n - then f(x) is guaranteed to be irreducible over F(pn).
By testing only some prime divisors q of n we can make the test probabilistic and reduce computation. If GCD results in 0, this indicates that f(x) shares a non-trivial factor with g(x), which means f(x) is not irreducible over F(pn).
Otherwise, brute force every possible divisor in F(2n), if none divide, then f(x) is irreducible over F(2n). p = x^(2^i), for every i from 1 to degree/2 mod p(x) congruent to x & p2 = p^(2^degree) mod p(x) congruent to x. Other method: Berlekamp's Algorithm
import { FieldPolynomial } from "polynomial-arithmetic"
const polynomial = new FieldPolynomial('x^7 + x^6 + 1')
console.log(polynomial.isPrimitive()) // trueA primitive polynomial (field theory) is monic, irreducible and has a root α that cyclically generates the entire field with αe = 1 thus having order e, it must have a non-zero constant term, for otherwise it will be divisible by x. Over GF(2), x + 1 is a primitive polynomial and all other primitive polynomials have an odd number of terms, since any polynomial mod 2 with an even number of terms is divisible by x + 1 (it has 1 as a root).
For a polynomial f(x) of degree m over GF(p) (where p is prime) to be primtive, the order which is the smallest positive integer e such that f(x) divides xe − 1 (e = pn − 1), must satisfy:
- xe ≡ 1 mod(f(x)) where e = 2n - 1 in GF(2)
- xd mod f(x) for each divisor d of (2^n - 1) - If some returns 1, f(x) is not primitive over GF(2n)
Other sources: Primitive Test, Primitive Roots, Cyclotomic Polynomial, Primitive Element Test
import { FieldPolynomial } from "polynomial-arithmetic"
// Generate High Degree LFSR Feddback shift register Polynomial
const gateTapsPoly = new FieldPolynomial("x^13 + x^12 + x^11 + x^8 + 1");
// Properties of a Maximum Length Sequence LFSR polynomial taps feed
const irreducible = gateTapsPoly.isIrreducible(); // true
const primitive = gateTapsPoly.isPrimitive(); // true
const setwise = gateTapsPoly.isSetwisePrime(); // true
console.log(`Maximum sequence LFSR Feedback polynomial: ${gateTapsPoly}`)Array of maximum sequence Linear Feedback Shift Registers (LFSRs) irreducible & primitive polynomials with setwise coprime taps. Following property names and their meaning:
- Degree: highest power of the variable 𝑥 with a non-zero coefficient in the array of polynomials
- Max: represents the maximum number of feedback polynomials to generate before stopping
- Possible: number of available polynomial tap combinations
- Generated: number of polynomial tap combinations with uneven terms
- Tested: number of polynomials that went through the test
- Found: number of polynomials that passed the test
This project encourages contributions and suggestions.