Monte Carlo methods require a series of random numbers. The most straightforward way to obtain these numbers is to use the rand() function in Matlab or a similar function in a different software system. These numbers are, of course, only pseudo-random – they are generated by an algorithm and no actual chance is involved. Section 13.1 in the book provides a discussion of some of the specific formulas used to generate pseudo-random numbers. The sequence produced by a good random number generator behaves like a sequence of true random numbers – that means that numbers must be uniformly distributed and there should be no way to predict the next number in the sequence from past results. Thus ideally, there should not be a difference whether pseudo random or true random numbers are used in a Monte Carlo method.

There are ways to generate large numbers of true random numbers efficiently from a physical source. For example here is a random number generator based on the quantum properties of photons. A photon projected on a semitransparent mirror can either be transmitted or reflected. Its behavior is completely random. This can be used to generate a true random sequence of zeros and ones and from there a sequence of random numbers. Random numbers can also be generated from a number of other sources such as circuit noise (used in the original Atari 8-bit computers), atmospheric noise , nuclear decay, and even by hashing images of the movement inside a lava lamp .

http://www.random.org/randomness/
http://en.wikipedia.org/wiki/Hardware_random_number_generator