期刊文献+
共找到1篇文章
< 1 >
每页显示 20 50 100
Greedy Algorithm Computing Minkowski Reduced Lattice Bases with Quadratic Bit Complexity of Input Vectors
1
作者 hao chen liqing xu 《Chinese Annals of Mathematics,Series B》 SCIE CSCD 2011年第6期857-862,共6页
The authors present an algorithm which is a modilication of the Nguyen-Stenle greedy reduction algorithm due to Nguyen and Stehle in 2009. This algorithm can be used to compute the Minkowski reduced lattice bases for ... The authors present an algorithm which is a modilication of the Nguyen-Stenle greedy reduction algorithm due to Nguyen and Stehle in 2009. This algorithm can be used to compute the Minkowski reduced lattice bases for arbitrary rank lattices with quadratic bit complexity on the size of the input vectors. The total bit complexity of the algorithm is O(n^2·(4n!)^n·(n!/2^n)^n/2·(4/3)^n(n-1)/2).log^2 A)where n is the rank of the lattice and A is maximal norm of the input base vectors. This is an O(log^2 A) algorithm which can be used to compute Minkowski reduced bases for the fixed rank lattices. A time complexity n!. 3n(log A)^O(1) algorithm which can be used to compute the successive minima with the help of the dual Hermite-Korkin-Zolotarev base was given by Blomer in 2000 and improved to the time complexity n!- (log A)^O(1) by Micciancio in 2008. The algorithm in this paper is more suitable for computing the Minkowski reduced bases of low rank lattices with very large base vector sizes. 展开更多
关键词 LATTICE Successive minima Minkowski reduced bases Greedy reduction
原文传递
上一页 1 下一页 到第
使用帮助 返回顶部