文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > Improved low-density subset sum algorithms
侵权投诉

Improved low-density subset sum algorithms

The general subset sum problem is NP-complete. However, there are two algorithms, one due to Brickell and the other to Lagarias and Odlyzko, which in polynomial time solve almost all subset sum problems of sufficiently low density. Both methods rely on bas

Therefore, Equation 6 can be replaced by: X

n

i=1 P=1+

xiai= y(t? 2s); 1 2

(13)

since ( n bi)? 2b0n 1= (0; 0;:::; 0; N (t? 2s)). i We now establish a bound on the size of jyj. From above,

jy(t? 2s)j= 2 If jt? 2sj 1 2

X

n

n n; where= max ai: i n 1

i=1

p

xiai (14) (15)

, then jyjjt? 2sj jyj, and

jyj 2npn;

by Equation 14. If jt? 2sj<, then we can solve two problems: one where is assumed to be part of the subset which sums to s, and one where is assumed to be part of the subset which sums to t? s. In the rst case, the new problem has s0= s?; t0= t?, and 1 2

jt0? 2s0j= jt?? 2s+ 2 j= jt? 2s+ j jt0? 2s0j= jt? 2s? j 1 2 1 2

1 2

:

(16) (17)

For the second case, the new problem has s0= s; t0= t?, and

:

and that the bound in Equation 15 holds. Thus we may always assume jt? 2sj We may now calculate the bound on probability P that there exists a vector x^ which satis es Equation 12. We now let x= (x;:::; xn) be any vector such that 2 x 2 Zn. We obtain the following bound, similar to Equation 10: 1

P Pr 1 2

X

n

i=1

aixi= y(t? 2s) 1 2=1 1 2

!

n

x: kxk

1 2

pn n 4npn+ 1: o

(18)

As in Section 2, Pr( n aixi= y(t? 2s)) 1=A. To estimate the number of vectors i pn, we again use the technique in 11, 14], but in a more complicated x with kxk way. The number of x with kxk pn=2 is bounded above by P n

w= (w;:::; wn): wi 2 Z for all i; kwk pn+ w= (w;:::; wn): wi 2 Z for all i; kw? (;;:::; )k 1 n 1 2 o 1 1 2 1 2 1 2

1 2

pn: (19) o

上一页第7页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email