|Authentication: We declare that the web-report is our own work and does not contain plagiarised material|
|Project Title||Ordered Randomness|
|Synopsis||We considered 2 of the PRNGs in popular programming languages, the Linear Congruential Generator (LCG) and the Subtract With Borrow (SWB) generator. We found that the period of each generator correlates with its performance on statistical tests such as the Diehard tests. The LCG was shown to have a much shorter period than the SWB and had a much worse performance on the Diehard tests than the SWB.
We then worked on making improvements to existing PRNGs by combining multiple PRNGs together into one larger PRNG. The two suggested methods are interseeding the PRNGs and combining PRNG values. The former maintains processing speed at the expense of memory and may not be able to reach its theoretical maximum period, while the latter allows for a significant increase in period but increases the amount of processing power needed and is an inefficient use of multiple PRNGs as only 1 value is obtained per iteration.
|Link to start page||Click HERE to access web-report|
evaluator to take note
|Best viewed in latest version of Google Chrome|
(Names & Classes)
|BENSON LIN ZHAN LI 3S1|
|LEE QIXIU, GALEN 3S1|
|LOH YAN XUN, TIMOTHY 3S1|