Bias in Monte Carlo Simulations Due To Pseudo-Random Number Generator Initial Seed Selection

Pseudo-random number generators can bias Monte Carlo simulations of the standard normal probability distribution function with initial seeds selection. Five generator designs were initial-seeded with values from 10000 HEX to 1FFFF HEX , estimates of the mean were calculated for each seed, the distribution of mean estimates was determined for each generator and simulation histories were graphed for selected seeds. Introduction It is possible, using a vetted pseudo-random number generator and a specific initial seed, to produce number sequences that may have very nonrandom-like characteristics; however, restarting the algorithm with a new seed will produce an excellent number sequence. Park and Miller (1988) asserted that coding errors in the algorithm might be responsible, although, in a vetted generator, coding errors are unlikely. (See Fishman (1995) for descriptions of other possible reasons for this phenomenon.)