Random Number and Its Performance Evaluation in Urban Traffic Simulation

Aiming at the disadvantage of linear congruential method widely used in urban traffic system simulation software,this paper adopts graphic approximating and statistical testing methods to compare the random number generators that being used and being researched.According to the norm theorem,it presents 2-norm statistical test performance evaluation expression and error.Analysis results indicate that MT algorithm can achieve better performance in traffic simulation.Example gives a proof that MT algorithm approximates actual traffic situation better.