扩展欧几里得算法就是求:
ax + by = gcd(a, b)
的一组整数解(x, y)
一、非递归的实现:
首先看a = 60, b = 22的情况:
表格左边是欧几里得算法,右边等式计算ax + by = gcd(a, b)的解
a = 2 × b + 16 | 16 = a - 2b |
b = 1 × 16 + 6 |
6 = b - 1 × 16 = b - 1 × (a - 2b) = -a + 3b |
16 = 2 × 6 + 4 |
4 = 16 - 2 × 6 = (a - 2b) - 2 × (-a + 3b) 3a - 8b |
6 = 1 × 4 + 2 |
2 = 6 - 1 × 4 = (-a + 3b) - 1 × (3a - 8b) = -4a + 11b |
4 = 2 × 2 + 0 |
这是《数论导引》里的伪代码:
那么x是如何计算出来的呢,x是上上次余数的a的系数减去这次求得的q乘以上次余数的a的系数
也就是r(n) = r(n-2) - q(n)×r(n-1) (括号中的数代表下标)
这句话有点绕,=_=!!。。
再看一般情况的表格:
a = q1b + r1 | r1 = a - q1b |
b = q2r1 + r2 | r2 = b - q2r1 |
r1 = q3r2 + r3 | r3 = r1 - q3r2 |
…… | …… |
1 int ex_gcd1(int a, int b, int &x, int &y) 2 { 3 int g, v, w, s, t, q; 4 x = 1; 5 v = 0; 6 g = a; 7 w = b; 8 while(w != 0) 9 { 10 q = g / w; 11 t = g % w; 12 s = x - q*v; 13 x = v; 14 g = w; 15 v = s; 16 w = t; 17 } 18 y = (g - a*x) / b; 19 return g; 20 }
二、递归方式的实现
令a‘ = a%b, t = a/b, y‘ = y + tx
ax + by = g
ax - tbx + tbx + by = g
(a - tb)x + b(tx + y) = g
a‘x + by‘ = g
by‘ + a‘x = g
y = y‘ - tx,y‘是比y深一度的递归的y
1 int ex_gcd2(int a, int b, int &x, int &y) 2 { 3 if(b == 0) 4 { 5 x = 1; 6 y = 0; 7 return a; 8 } 9 int g = ex_gcd2(b, a%b, y, x); 10 y = y - a/b*x; 11 return g; 12 }
原文:http://www.cnblogs.com/AOQNRMGYXLMV/p/3856173.html