The java.util.Random class allows you to create
objects that produce pseudo-random numbers with uniform or gaussian
distributions according to a linear congruential formula with a
48-bit seed.
The algorithm used is good enough for single-player, no-money games. I wouldn't use it for cryptography.
You can choose the seed or you can let Java pick one based on the current time.
Random r = new Random(109876L);
int i = r.nextInt();
int j = r.nextInt();
long l = r.nextLong();
float f = r.nextFloat();
double d = r.nextDouble();
int k = r.nextGaussian();
The nextInt(), nextLong(), and
nextBytes() methods all cover their respective ranges with
equal likelihood. For example, to simulate a six-sided die; that is
to generate a random integer between 1 and 6, you might write
Random r = new Random();
int die = r.nextInt();
die = Math.abs(die);
die = die % 6;
die += 1;
System.out.println(die);
The nextGaussian() method returns a pseudo-random,
Gaussian distributed, double value with mean 0.0 and standard
deviation 1.0.
The nextBytes() method fills a byte[] array with
random bytes. For example,
byte[] ba = new byte[1024];
Random r = new Random();
r.nextBytes(ba);
for (int i = 0; i < ba.length; i++) {
System.out.println(ba[i]);
}