There are two groups of numbers: predictable and random. Random numbers are the fundamental for many applications such as data security, statistical modeling and operational research. There are numerous random number generators available and selection of an appropriate right random number generator is very important. This is due to the fact that using the wrong generator could have serious consequences to the wellness efficiency and security of the operation.

Therefore, in this paper, we implement six most popular pseudo-random number generators using MATLAB software and compare the randomness behavior of the generated numbers in terms of their uniformity, dependency and execution speed. Initially, the pseudo-random numbers are examined for uniform distribution using the chi-square test. In addition, the numbers are then examined for dependency behavior using a serial test. Finally, the number of iterations is used as a parameter in order to determine their execution speeds. Based on our simulation, the results showed that shr3cong appears to outperform other generators in which it consistently abiding the confidence interval with an exception of failing in only one single entry on the three-dimensional test. In terms of execution speed, mcg16807 becomes the fastest algorithm which outclassed all the others.

chi-square test, serial test, independence, uniform distribution, code timing.