Sixcarbs
Active Member
I just finished reading Brooks and watching David's video on Monte Carlo simulations.
The age old problem of computer generated RNG's came up as not being truly random because they always require a seed.
What we used to do when I had a commercial venture that required a true random numbers was introduce some external physical numbers to set the seed. I don't remember what we finally settled on but the possibilities included, instantaneous outside temperatures of different cities, cpu temperatures, processor loads, and the seconds on the clock at the exact moment the RNG was started.
By using a product of real time external and independent event numbers we generated a different and random seed each time.
The age old problem of computer generated RNG's came up as not being truly random because they always require a seed.
What we used to do when I had a commercial venture that required a true random numbers was introduce some external physical numbers to set the seed. I don't remember what we finally settled on but the possibilities included, instantaneous outside temperatures of different cities, cpu temperatures, processor loads, and the seconds on the clock at the exact moment the RNG was started.
By using a product of real time external and independent event numbers we generated a different and random seed each time.
