Random number generator: Difference between revisions
(Link to rng (what).) |
No edit summary |
||
| (One intermediate revision by one other user not shown) | |||
| Line 1: | Line 1: | ||
A '''random number generator''' picks numbers at chance from a distribution. |
A '''random number generator''' picks numbers at chance from a distribution. |
||
The most common type is a ''uniform generator'', where each number has an equal chance. If a uniform generator yields decimal digits from 0 to 9, then each digit has a 10% chance. The other type is a [[unbias a random generator|''weighted'' or ''biased'' generator]], where the chances are not equal. A biased generator that yields 0 with 70% chance, yields 1 with 30% chance, would probably yield more zeros than ones. |
|||
''True random numbers'' are impossible to predict. Many programs use ''pseudorandom numbers'', which are not as good. A pseudorandom generator uses a formula and a state to calculate a sequence of numbers. Anyone who knows the formula and the state can predict those numbers. |
|||
Pseudorandom numbers can also have statistical problems. True random numbers are [[wp:independence (probability theory)|independent]]. A pseudorandom sequence may contain patterns where the values of some numbers change the probabilities of other numbers. |
|||
See also [[Pick_random_element]]. |
|||
== Tasks == |
|||
* [[Random number generator (included)]] |
* [[Random number generator (included)]] |
||
* [[Random number generator (device)]] |
* [[Random number generator (device)]] |
||
* Everything in [[:Category:Randomness]] |
|||
== Links == |
|||
* Wikipedia's article about [[wp:Random number generation|Random number generation]] |
|||
* [http://csrc.nist.gov/groups/ST/toolkit/rng/index.html NIST's statistical tests for random numbers] |
|||
[[Category:Encyclopedia]] |
[[Category:Encyclopedia]] |
||
Latest revision as of 15:19, 9 August 2011
A random number generator picks numbers at chance from a distribution.
The most common type is a uniform generator, where each number has an equal chance. If a uniform generator yields decimal digits from 0 to 9, then each digit has a 10% chance. The other type is a weighted or biased generator, where the chances are not equal. A biased generator that yields 0 with 70% chance, yields 1 with 30% chance, would probably yield more zeros than ones.
True random numbers are impossible to predict. Many programs use pseudorandom numbers, which are not as good. A pseudorandom generator uses a formula and a state to calculate a sequence of numbers. Anyone who knows the formula and the state can predict those numbers.
Pseudorandom numbers can also have statistical problems. True random numbers are independent. A pseudorandom sequence may contain patterns where the values of some numbers change the probabilities of other numbers.
See also Pick_random_element.
Tasks
- Random number generator (included)
- Random number generator (device)
- Everything in Category:Randomness
Links
- Wikipedia's article about Random number generation
- NIST's statistical tests for random numbers