Structure and interpretation of computer programs
Structure and interpretation of computer programs
Efficient and portable combined random number generators
Communications of the ACM
The art of computer programming, volume 2 (3rd ed.): seminumerical algorithms
The art of computer programming, volume 2 (3rd ed.): seminumerical algorithms
| Hi-index | 0.00 |
Computer simulations are everywhere, from the corporate office to the local video game parlor. With the increased role being played by these simulations, it is important for students to be completely aware of the limitations of pseudorandom number generators. The fact that random number generators in use today are not truly random is no secret (The New York Times, C1-C8). Since most simulations produce reasonable results it might be difficult to convince students that there are any problems involved in trusting these random number generators. A simple simulation which can be used as a programming exercise in any language can dramatically reveal these dangers.The exercise used requires a statistical evaluation of π which yields horrible results. The results are not difficult to explain and the exercise can be extended by allowing students to experiment with modifications to the pseudo-random number generator used in attempts to "fix" the problem. This can be useful, as the attempts to fix the pseudorandom number generator usually aggravate rather than alleviate the problem.