#RANDOM SEQUENCE GENERATOR GENERATOR#

This variable takes a comma-separated list of lengths. The groupLength variable controls how many bits will appear in a single bit group.

Otherwise, groups of multiple bits will be generated. If the group length is one, then single bits will be generated. In every iteration, it creates an empty array groupBits = that will store the randomly generated bits. Generating individual bits is the same as generating bit groups of length one. This variable can be set in the options and it determines how many individual bits or bit groups will appear in the output. To generate random bits, it runs two nested for loops.

#RANDOM SEQUENCE GENERATOR GENERATOR#

One thing that comes to mind is that if your range is 0.This random bit generator works entirely in your browser and is written in JavaScript. The key is that the more random bits you infuse the harder it is to reverse engineer. Some combination of these should give you a hard to reverse engineer sequence. Generate a random permutation over log_2(N) bits at the beginning of sequence using the above method and then swap the bits in the result above.įind a random number r1 that is a relative prime to n at and another random number r2 between 0 and n-1. keep a counter i for the sequence that goes from 0 through N-1.ĭo log_2(N/2) bit flips on i where bits to flip are chosen at random when when you are starting the sequence. Now since you want these many random bits stored anywhere I think (more like a gut feeling than a mathematical proof) it would be hard to do a truly random permutation without that much storage).īut if you just want a sequence that is not easy to reverse engineer you could just concatenate a number of 1:1 functions one after the other. So you do need that many random bits to be generated. Note that the inherent information in a randomly chosen permutation is log(n!) which is somewhere close to n/e bits. How random do you want the sequence to be. This is in response to the comments below: I'd suggest starting with an array with A=i It is quite possible to get something wrong. Unless you have good reasons to, you don't want to reinvent pseudo random generation. UPDATE: As I posted in the comment here, there's already a generator like this: AES Encryption, but unfortunately it can only be used for 128-bit ranges.

I'm aware and used once the suites that can test general random sequences, but this one is a different so probably some special suite needed or at least some conversion back to the general random world Is there an algorithm or a test suite for testing such sequences.For example 200 operations for numbers from 0 to 31 gives visually good results, but 200 operations for the range 0.199 sometimes gives blocks of close numbers. How can I approximately guess the number of operations for a given range in order for the sequence to be good enough.Is there smallest subset of the operations that will be enough (for example invert bit + xor bit by another bit) and adding any other just will make the algorithm harder to read while giving no extra benefits.Now the sequences looks much better, but the questions are: The real game changer was xor one bit by another bit. If we use only ones mentioned, the quality of the sequence won't be great since the nearby counter will produce results with similar number of zeros and ones. There are at least several operation possible like exchange two bits, invert a bit, cyclic shift. Since the operations are reversible, different source numbers will give different result numbers. I read a couple of posts here about generating random sequence without repeats (for example Create Random Number Sequence with No Repeats) and decided to implement it for my own needĪctually it was an algorithm applying some non-destructive (reversible) operations with the bits of the current counter in order to get a pseudo-random number that should appear only once.

Random sequence generator