Class Random

public Random()

Creates a new pseudorandom number generator. The seed is initialized to the current time, as if by setSeed(System.currentTimeMillis());.

See Also:

System.currentTimeMillis()

public Random(long seed)

Creates a new pseudorandom number generator, starting with the specified seed, using setSeed(seed);.

Parameters:

seed - the initial seed

protected int next(int bits)

Generates the next pseudorandom number. This returns an int value whose bits low order bits are independent chosen random bits (0 and 1 are equally likely). The implementation for java.util.Random is:

protected synchronized int next(int bits)
{
seed = (seed * 0x5DEECE66DL + 0xBL) & ((1L << 48) - 1);
return (int) (seed >>> (48 - bits));
}

Parameters:

bits - the number of random bits to generate, in the range 1..32

Returns:

the next pseudorandom value

Since:

1.1

public boolean nextBoolean()

Generates the next pseudorandom boolean. True and false have the same probability. The implementation is:

public boolean nextBoolean()
{
return next(1) != 0;
}

Returns:

the next pseudorandom boolean

Since:

1.2

public void nextBytes(byte[] bytes)

Fills an array of bytes with random numbers. All possible values are (approximately) equally likely. The JDK documentation gives no implementation, but it seems to be:

public void nextBytes(byte[] bytes)
{
for (int i = 0; i < bytes.length; i += 4)
{
int random = next(32);
for (int j = 0; i + j < bytes.length && j < 4; j++)
{
bytes[i+j] = (byte) (random & 0xff)
random >>= 8;
}
}
}

Parameters:

bytes - the byte array that should be filled

Throws:

NullPointerException - if bytes is null

Since:

1.1

public double nextDouble()

Generates the next pseudorandom double uniformly distributed between 0.0 (inclusive) and 1.0 (exclusive). The implementation is as follows.

public double nextDouble()
{
return (((long) next(26) << 27) + next(27)) / (double)(1L << 53);
}

Returns:

the next pseudorandom double

public float nextFloat()

Generates the next pseudorandom float uniformly distributed between 0.0f (inclusive) and 1.0f (exclusive). The implementation is as follows.

public float nextFloat()
{
return next(24) / ((float)(1 << 24));
}

Returns:

the next pseudorandom float

public double nextGaussian()

Generates the next pseudorandom, Gaussian (normally) distributed double value, with mean 0.0 and standard deviation 1.0. The algorithm is as follows.

public synchronized double nextGaussian()
{
if (haveNextNextGaussian)
{
haveNextNextGaussian = false;
return nextNextGaussian;
}
else
{
double v1, v2, s;
do
{
v1 = 2 * nextDouble() - 1; // between -1.0 and 1.0
v2 = 2 * nextDouble() - 1; // between -1.0 and 1.0
s = v1 * v1 + v2 * v2;
}
while (s >= 1);
double norm = Math.sqrt(-2 * Math.log(s) / s);
nextNextGaussian = v2 * norm;
haveNextNextGaussian = true;
return v1 * norm;
}
}

This is described in section 3.4.1 of The Art of Computer Programming, Volume 2 by Donald Knuth.

Returns:

the next pseudorandom Gaussian distributed double

public int nextInt()

Generates the next pseudorandom number. This returns an int value whose 32 bits are independent chosen random bits (0 and 1 are equally likely). The implementation for java.util.Random is:

public int nextInt()
{
return next(32);
}

Returns:

the next pseudorandom value

public int nextInt(int n)

Generates the next pseudorandom number. This returns a value between 0(inclusive) and n(exclusive), and each value has the same likelihodd (1/n). (0 and 1 are equally likely). The implementation for java.util.Random is:

public int nextInt(int n)
{
if (n <= 0)
throw new IllegalArgumentException("n must be positive");
if ((n & -n) == n)  // i.e., n is a power of 2
return (int)((n * (long) next(31)) >> 31);
int bits, val;
do
{
bits = next(31);
val = bits % n;
}
while(bits - val + (n-1) < 0);
return val;
}

This algorithm would return every value with exactly the same probability, if the next()-method would be a perfect random number generator. The loop at the bottom only accepts a value, if the random number was between 0 and the highest number less then 1<<31, which is divisible by n. The probability for this is high for small n, and the worst case is 1/2 (for n=(1<<30)+1). The special treatment for n = power of 2, selects the high bits of the random number (the loop at the bottom would select the low order bits). This is done, because the low order bits of linear congruential number generators (like the one used in this class) are known to be ``less random'' than the high order bits.

Parameters:

n - the upper bound

Returns:

the next pseudorandom value

Throws:

IllegalArgumentException - if the given upper bound is negative

Since:

1.2

public long nextLong()

Generates the next pseudorandom long number. All bits of this long are independently chosen and 0 and 1 have equal likelihood. The implementation for java.util.Random is:

public long nextLong()
{
return ((long) next(32) << 32) + next(32);
}

Returns:

the next pseudorandom value

public void setSeed(long seed)

Sets the seed for this pseudorandom number generator. As described above, two instances of the same random class, starting with the same seed, should produce the same results, if the same methods are called. The implementation for java.util.Random is:

public synchronized void setSeed(long seed)
{
this.seed = (seed ^ 0x5DEECE66DL) & ((1L << 48) - 1);
haveNextNextGaussian = false;
}

Parameters:

seed - the new seed