Korkine–Zolotarev lattice basis reduction algorithm explained

Korkine–Zolotarev lattice basis reduction algorithm should not be confused with Kolmogorov–Zurbenko filter.

The Korkine–Zolotarev (KZ) lattice basis reduction algorithm or Hermite–Korkine–Zolotarev (HKZ) algorithm is a lattice reduction algorithm.

For lattices in

Rⁿ

it yields a lattice basis with orthogonality defect at most

nⁿ

, unlike the

	n^2/2
2

bound of the LLL reduction.^[1] KZ has exponential complexity versus the polynomial complexity of the LLL reduction algorithm, however it may still be preferred for solving multiple closest vector problems (CVPs) in the same lattice, where it can be more efficient.

History

The definition of a KZ-reduced basis was given by Aleksandr Korkin and Yegor Ivanovich Zolotarev in 1877, a strengthened version of Hermite reduction. The first algorithm for constructing a KZ-reduced basis was given in 1983 by Kannan.^[2]

The block Korkine-Zolotarev (BKZ) algorithm was introduced in 1987.^[3]

Definition

A KZ-reduced basis for a lattice is defined as follows:^[4]

Given a basis

B=\{b_1,b_2,...,b_n\},

define its Gram–Schmidt process orthogonal basis

B^*=\{

	*
b
	1,

	*
b
	2,

...,

	*
b
	n

\},

and the Gram-Schmidt coefficients

\mu_i,j=

	*
\langleb
	j\rangle

	*
\langleb
	j\rangle

, for any

1\lej<i\len

Also define projection functions

\pi_i(x)=\sum_j

	*
\langlex,b
	j\rangle

	*
\langleb
	j\rangle

	*
b
	j

which project

orthogonally onto the span of

	*
b
	i,

… ,

	*
b
	n

Then the basis

is KZ-reduced if the following holds:

	*
b
	i

is the shortest nonzero vector in

\pi_i(l{L}(B))

For all

j<i

\left|\mu_i,j\right|\leq1/2

Note that the first condition can be reformulated recursively as stating that

b₁

is a shortest vector in the lattice, and

\{\pi_2(b_2), … \pi_2(b_n)\}

is a KZ-reduced basis for the lattice

\pi_2(l{L}(B))

Also note that the second condition guarantees that the reduced basis is length-reduced (adding an integer multiple of one basis vector to another will not decrease its length); the same condition is used in the LLL reduction.

References

A. . Korkine. G.. Zolotareff. Sur les formes quadratiques positives. Mathematische Annalen. 1877. 11. 2. 242–292. 121803621. 10.1007/BF01442667.
Shanxiang . Lyu. Cong. Ling. Boosted KZ and LLL Algorithms. IEEE Transactions on Signal Processing. 2017. 65. 18. 4784–4796. 10.1109/TSP.2017.2708020. 1703.03303. 2017ITSP...65.4784L. 16832357.
Jinming . Wen. Xiao-Wen. Chang. On the KZ Reduction. 2018. 1702.08152.
Book: Daniele. Micciancio. Shafi. Goldwasser. Complexity of Lattice Problems. 2002. 131–136. 10.1007/978-1-4615-0897-7. 978-1-4613-5293-8.
Wen. Zhang. Sanzheng. Qiao. Yimin. Wei. HKZ and Minkowski Reduction Algorithms for Lattice-Reduction-Aided MIMO Detection. IEEE Transactions on Signal Processing . 2012. 60 . 11 . 5963 . 10.1109/TSP.2012.2210708 . 2012ITSP...60.5963Z . 5962834 .

Notes and References

https://sites.math.washington.edu/~rothvoss/lecturenotes/IntOpt-and-Lattices.pdf, pp. 18-19
Zhang et al 2012, p.1
Book: https://link.springer.com/chapter/10.1007/978-981-15-5191-8_15. 10.1007/978-981-15-5191-8_15. A Survey of Solving SVP Algorithms and Recent Strategies for Solving the SVP Challenge. International Symposium on Mathematics, Quantum Theory, and Cryptography. Mathematics for Industry. 2021. Yasuda. Masaya. 33. 189–207. 978-981-15-5190-1. 226333525.
Micciancio & Goldwasser, p.133, definition 7.8