Random Number Generators

We define a sequence of numbers in the interval as being random if there are no correlations between different numbers in the sequence. Yet randomness does not necessarily mean all numbers in the sequence are all equally likely to occur. They are random with the distribution if the probability of finding in the interval is . The standard random number generator on computers generates uniform distributions (P =1) between 0 and 1. In other words there should be an equal probability of generating any number in this interval, with the next number generated independently of the previous number.

Computers are deterministic. They cannot create a truly random sequence in which we cannot predict even if we know all preceding elements - it may be a bit of work for us to actually sit down and do the figuring, but it is always possible. For this reason computers generate what are called ``pseudo'' random numbers. By nature of their creation, essentially all computed random numbers must contain correlations and thus are not truly random. (With our incurable laziness we won't bother saying pseudo all the time.) While the more sophisticated of the generators do a better job at hiding the correlations, experience shows that if you look hard enough or use these numbers enough you will notice correlations. An alternative to generating random numbers is to read in a table of true random numbers, that is, numbers determined by naturally random processes such as radioactive decay. If you need follow such an approach, you will find the needed tables in literature.

• Method: Linear Congruent Method
• Project