int randbit( void )
{
    while( 1 )
    {
        int r = rand5();
        if( r <= 4 ) return(r & 1);
    }
}

int randint( int nbits )
{
    int result = 0;
    while( nbits-- )
    {
        result = (result<<1) | randbit();
    }
    return( result );
}

int rand7( void )
{
    while( 1 )
    {
        int r = randint( 3 ) + 1;
        if( r <= 7 ) return( r );
    }
}

下面使用随机数发生器在{1,2,3,4,5,6,7}上产生均匀分布，在{1,2,3,4,5}上产生均匀分布。代码很混乱，但逻辑很清晰。

public static int random_7(Random rg) {
    int returnValue = 0;
    while (returnValue == 0) {
        for (int i = 1; i <= 3; i++) {
            returnValue = (returnValue << 1) + SimulateFairCoin(rg);
        }
    }
    return returnValue;
}

private static int SimulateFairCoin(Random rg) {
    while (true) {
        int flipOne = random_5_mod_2(rg);
        int flipTwo = random_5_mod_2(rg);

        if (flipOne == 0 && flipTwo == 1) {
            return 0;
        }
        else if (flipOne == 1 && flipTwo == 0) {
            return 1;
        }
    }
}

private static int random_5_mod_2(Random rg) {
    return random_5(rg) % 2;
}

private static int random_5(Random rg) {
    return rg.Next(5) + 1;
}    

什么是简单的解决方案?(rand5() + rand5()) % 7 + 1 减少内存使用或在较慢的CPU上运行的有效解决方案是什么?是的，这是有效的，因为它只调用rand5()两次，空间复杂度为O(1)

考虑rand5()给出从1到5(包括)的随机数。 (1 + 1) % 7 + 1 = 3 (1 + 2) % 7 + 1 = 4 (1 + 3) % 7 + 1 = 5 (1 + 4) % 7 + 1 = 6 (1 + 5) % 7 + 1 = 7

(2 + 1) % 7 + 1 = 4 (2 + 2) % 7 + 1 = 5 (2 + 3) % 7 + 1 = 6 (2 + 4) % 7 + 1 = 7 (2 + 5) % 7 + 1 = 1 .

(5 + 1) % 7 + 1 = 7 (5 + 2) % 7 + 1 = 1 (5 + 3) % 7 + 1 = 2 (5 + 4) % 7 + 1 = 3 (5 + 5) % 7 + 1 = 4 .

等等

因为1/7是一个以5为底的无限小数，所以没有(完全正确的)解可以在常数时间内运行。一个简单的解决方案是使用拒绝抽样，例如:

int i;
do
{
  i = 5 * (rand5() - 1) + rand5();  // i is now uniformly random between 1 and 25
} while(i > 21);
// i is now uniformly random between 1 and 21
return i % 7 + 1;  // result is now uniformly random between 1 and 7

这个循环的预期运行时间为25/21 = 1.19次迭代，但是永远循环的概率非常小。

假设rand(n)在这里表示“从0到n-1均匀分布的随机整数”，下面是使用Python的randint的代码示例，它具有这种效果。它只使用randint(5)和常量来产生randint(7)的效果。其实有点傻

from random import randint
sum = 7
while sum >= 7:
    first = randint(0,5)   
    toadd = 9999
    while toadd>1:
        toadd = randint(0,5)
    if toadd:
        sum = first+5
    else:
        sum = first

assert 7>sum>=0 
print sum

这里是我的一般实现，在给定一个范围为[0,B-1]的均匀发生器的情况下，生成范围为[0,N-1]的均匀。

public class RandomUnif {

    public static final int BASE_NUMBER = 5;

    private static Random rand = new Random();

    /** given generator, returns uniform integer in the range 0.. BASE_NUMBER-1
    public static int randomBASE() {
        return rand.nextInt(BASE_NUMBER);
    }

    /** returns uniform integer in the range 0..n-1 using randomBASE() */
    public static int randomUnif(int n) {
        int rand, factor;
        if( n <= 1 ) return 0;
        else if( n == BASE_NUMBER ) return randomBASE();
        if( n < BASE_NUMBER ) {
            factor = BASE_NUMBER / n;
            do
                rand = randomBASE() / factor;
            while(rand >= n);
            return rand;
        } else {
            factor = (n - 1) / BASE_NUMBER + 1;
            do {
                rand = factor * randomBASE() + randomUnif(factor);
            } while(rand >= n);
            return rand;
        }
    }
}

不是特别高效，但一般和紧凑。对基生成器的均值调用:

 n  calls
 2  1.250 
 3  1.644 
 4  1.252 
 5  1.000 
 6  3.763 
 7  3.185 
 8  2.821 
 9  2.495 
10  2.250 
11  3.646 
12  3.316 
13  3.060 
14  2.853 
15  2.650 
16  2.814 
17  2.644 
18  2.502 
19  2.361 
20  2.248 
21  2.382 
22  2.277 
23  2.175 
24  2.082 
25  2.000 
26  5.472 
27  5.280 
28  5.119 
29  4.899