Java Random Numbers

Java Random Numbers - Random Numbers With Math.random()

• Computers Don't Generate Truly Random Numbers

  Computers cannot create true random numbers according to strict mathematical definitions but for most general computing needs this is not a practical limitation. The pseudorandom numbers generated by computers will be a close enough approximation. (If it is not sufficient for your purposes, consult a mathematician!)

• Random Number Generation Example

  To start, let's look at a simple example.

   double randomValue = Math.random();  

  What did you get? I guarantee that it was a number between 0.0 and 1.0 (including 0.0 but not including 1.0). I got "0.8416533062276157." Run it again. You probably got a different number. Pretty easy, right?

  So, the default behavior of Math.random() is to generate a double in the interval [0 1).

• Build a Custom Random Number Generator

  Generating numbers in the range 0.0 to 1.0 has limited value, so let's see how we can define the range that we'll pull our random number from the range 0.0 to 10.0 (including 0.0 but not including 10.0) or [0 10). We do this by multiplying the generated random double by 10. This is called scaling the range.

   double randomValue = Math.random() * 10;  

  I got "5.796280106258913" when I ran this example.

  So, let's change the interval to be 1.0 to 11.0 [1 11). This is known as shifting the range.

   double randomValue = 1 + Math.random() * 10;  

  Running this example, I got "7.6887892451197315." It is hard to see any difference, so let's prove that we are shifting the range. Let's generate a random number from 100.0 to 200.0 or [100 200).

   double randomValue = 100 + Math.random() * 100;  

  I got 149.1804870735694. (I won't say that I told you so.)

• Generate Random Integers

  If one actually needs an int instead of a double, simply convert the returned double to an int by casting it.

  int randomInt = 1 + (int)(Math.random() * 50);  

  This will truncate the generated double and return an int between 1 and 50. I got "19." Don't forget the parentheses around the Math.random() * 50. This is required, otherwise the double will be truncated before the multiplication operation!

Java Random Numbers With Random Class

Occasionally, you may need more help generating random numbers than Math.random() can give. In these cases, you can turn to java.util.Random

Seeding Random Number Generation

So far, we have been asking for random numbers and expecting them to be truly random (or at least as random as a computer can produce). This works well except when we are testing and debugging code that manipulates randomly generated numbers.

For instance, if you are writing a method that "rolls" two 6-sided dice and returns the resulting sum, how do you know if your calculation is correct? Each time you call the method, each die will return a random number. For this and other scenarios where you want random number generation to always return the same sequence, you can seed the generator with a long.

Let's take a look at generating random numbers with Random and how we can define the seed for testing and debugging purposes.

 Random randomizer = new Random(1L); //remember longs are appended with L  

First, we create a new Random instance and pass in our desired seed value of 1L. Any two instances of Random that have the same seed will always return the same "random" numbers in the exact same sequence.

 int randomInt = randomizer.nextInt();  

Next, we generate our random number, randomInt. Notice we haven't defined an interval, as we did with Math.random(). Random.nextInt() returns a number between (approximately) -2.1 billion and 2.1 billion. (This is based on the size an int can hold.)

randomInt = Math.abs(randomInt);  

Next, we pass randomInt through Math.abs(). This, of course, eliminates negatives.

 randomInt = (randomInt % 6) + 1;  

This step looks similar to, but not exactly like, the scaling and shifting we did with Math.random(). The idea is the same, we want to customize our random number's range (from 2.1 billion to 6). So, we first find randomInt modulo 6. This gives us a number between 0 and 5. Finally, we shift the result up 1, giving us a final range of 1 to 6. Time to take a break!

Random Class Wrap-Up

When you are not debugging or testing code, you can randomly seed your random number generator by passing in the long from System.currentTimeMillis(). This grabs a long from the system's clock.

As you saw in the above examples, using Math.random() is significantly simplier than using Random. Scenarios when Random is helpful:

• One needs to seed the random number sequence.
• One wants to easily get a random byte.
• One needs a random number from a Gaussian distribution.

Now you are properly equipped to whip up virtually any random number you may need in your coding projects. Let's put your knowledge of random number generation to use by writing a set of classes to simulate rolling dice.