I am trying to understand the following PRNG algorithm patented by IBM:
The present invention relates to the field of computer data encrypting and decrypting, especially for mobile equipments like PDA, mobile phones, smart cards and the like, which need a good trade-off between computing speed, power consumption and security strength. Embodiments of the invention provide encrypting/decrypting methods implementing simple data operation. Such methods are based on generating a pseudo-random sequence through a function of the Collatz (or Syracuse) family from a starting number used as a secret key.
This algorithm is about generating pseudo-random numbers with Collatz sequences and generalized Collatz functions. Unfortunately, it is written in such an unclear way that I don't understand almost anything.
Is someone able to explain in a clear way what the presented algorithm does? I am pretty sure it can be described in a few lines, but for some reasons the authors did everything to make it obscure.