Berlekamp–Rabin algorithm

In number theory, Berlekamp's root finding algorithm, also called the Berlekamp–Rabin algorithm, is the probabilistic method of finding roots of polynomials over the field with elements. The method was discovered by Elwyn Berlekamp in 1970[1] as an auxiliary to the algorithm for polynomial factorization over finite fields. The algorithm was later modified by Rabin for arbitrary finite fields in 1979.[2] The method was also independently discovered before Berlekamp by other researchers.[3]

Elwyn R. Berlekamp at conference on Combinatorial Game Theory at Banff International Research Station Elwyn Berlekamp

History

edit

The method was proposed by Elwyn Berlekamp in his 1970 work[1] on polynomial factorization over finite fields. His original work lacked a formal correctness proof[2] and was later refined and modified for arbitrary finite fields by Michael Rabin.[2] In 1986 René Peralta proposed a similar algorithm[4] for finding square roots in  .[5] In 2000 Peralta's method was generalized for cubic equations.[6]

Statement of problem

edit

Let   be an odd prime number. Consider the polynomial   over the field   of remainders modulo  . The algorithm should find all   in   such that   in  .[2][7]

Algorithm

edit

Randomization

edit

Let  . Finding all roots of this polynomial is equivalent to finding its factorization into linear factors. To find such factorization it is sufficient to split the polynomial into any two non-trivial divisors and factorize them recursively. To do this, consider the polynomial   where   is some element of  . If one can represent this polynomial as the product   then in terms of the initial polynomial it means that  , which provides needed factorization of  .[1][7]

Classification of   elements

edit

Due to Euler's criterion, for every monomial   exactly one of following properties holds:[1]

  1. The monomial is equal to   if  ,
  2. The monomial divides   if   is quadratic residue modulo  ,
  3. The monomial divides   if   is quadratic non-residual modulo  .

Thus if   is not divisible by  , which may be checked separately, then   is equal to the product of greatest common divisors   and  .[7]

Berlekamp's method

edit

The property above leads to the following algorithm:[1]

  1. Explicitly calculate coefficients of  ,
  2. Calculate remainders of   modulo   by squaring the current polynomial and taking remainder modulo  ,
  3. Using exponentiation by squaring and polynomials calculated on the previous steps calculate the remainder of   modulo  ,
  4. If   then   mentioned below provide a non-trivial factorization of  ,
  5. Otherwise all roots of   are either residues or non-residues simultaneously and one has to choose another  .

If   is divisible by some non-linear primitive polynomial   over   then when calculating   with   and   one will obtain a non-trivial factorization of  , thus algorithm allows to find all roots of arbitrary polynomials over  .

Modular square root

edit

Consider equation   having elements   and   as its roots. Solution of this equation is equivalent to factorization of polynomial   over  . In this particular case problem it is sufficient to calculate only  . For this polynomial exactly one of the following properties will hold:

  1. GCD is equal to   which means that   and   are both quadratic non-residues,
  2. GCD is equal to  which means that both numbers are quadratic residues,
  3. GCD is equal to  which means that exactly one of these numbers is quadratic residue.

In the third case GCD is equal to either   or  . It allows to write the solution as  .[1]

Example

edit

Assume we need to solve the equation  . For this we need to factorize  . Consider some possible values of  :

  1. Let  . Then  , thus  . Both numbers   are quadratic non-residues, so we need to take some other  .
  1. Let  . Then  , thus  . From this follows  , so   and  .

A manual check shows that, indeed,   and  .

Correctness proof

edit

The algorithm finds factorization of   in all cases except for ones when all numbers   are quadratic residues or non-residues simultaneously. According to theory of cyclotomy,[8] the probability of such an event for the case when   are all residues or non-residues simultaneously (that is, when   would fail) may be estimated as   where   is the number of distinct values in  .[1] In this way even for the worst case of   and  , the probability of error may be estimated as   and for modular square root case error probability is at most  .

Complexity

edit

Let a polynomial have degree  . We derive the algorithm's complexity as follows:

  1. Due to the binomial theorem  , we may transition from   to   in   time.
  2. Polynomial multiplication and taking remainder of one polynomial modulo another one may be done in  , thus calculation of   is done in  .
  3. Binary exponentiation works in  .
  4. Taking the   of two polynomials via Euclidean algorithm works in  .

Thus the whole procedure may be done in  . Using the fast Fourier transform and Half-GCD algorithm,[9] the algorithm's complexity may be improved to  . For the modular square root case, the degree is  , thus the whole complexity of algorithm in such case is bounded by   per iteration.[7]

References

edit
  1. ^ a b c d e f g Berlekamp, E. R. (1970). "Factoring polynomials over large finite fields". Mathematics of Computation. 24 (111): 713–735. doi:10.1090/S0025-5718-1970-0276200-X. ISSN 0025-5718.
  2. ^ a b c d M. Rabin (1980). "Probabilistic Algorithms in Finite Fields". SIAM Journal on Computing. 9 (2): 273–280. CiteSeerX 10.1.1.17.5653. doi:10.1137/0209024. ISSN 0097-5397.
  3. ^ Donald E Knuth (1998). The art of computer programming. Vol. 2 Vol. 2. ISBN 978-0201896848. OCLC 900627019.
  4. ^ Tsz-Wo Sze (2011). "On taking square roots without quadratic nonresidues over finite fields". Mathematics of Computation. 80 (275): 1797–1811. arXiv:0812.2591. doi:10.1090/s0025-5718-2011-02419-1. ISSN 0025-5718. S2CID 10249895.
  5. ^ R. Peralta (November 1986). "A simple and fast probabilistic algorithm for computing square roots modulo a prime number (Corresp.)". IEEE Transactions on Information Theory. 32 (6): 846–847. doi:10.1109/TIT.1986.1057236. ISSN 0018-9448.
  6. ^ C Padró, G Sáez (August 2002). "Taking cube roots in Zm". Applied Mathematics Letters. 15 (6): 703–708. doi:10.1016/s0893-9659(02)00031-9. ISSN 0893-9659.
  7. ^ a b c d Alfred J. Menezes, Ian F. Blake, XuHong Gao, Ronald C. Mullin, Scott A. Vanstone (1993). Applications of Finite Fields. The Springer International Series in Engineering and Computer Science. Springer US. ISBN 9780792392828.{{cite book}}: CS1 maint: multiple names: authors list (link)
  8. ^ Marshall Hall (1998). Combinatorial Theory. John Wiley & Sons. ISBN 9780471315186.
  9. ^ Aho, Alfred V. (1974). The design and analysis of computer algorithms. Addison-Wesley Pub. Co. ISBN 0201000296.