Stochastic seeds permit reproducing a set of random-numbers. For getting a new set of random-numbers you just have to change one value: the seed. If you reuse a seed, you'll get the same random-numbers again.
- The stochastic seed is a basic value for all stochastic distributed values in the simulation from where the random numbers are generated.
- Running the same experiment twice without changing any parameters will result in two identical results.
- By changing the stochastic seed between two simulation runs (and by keeping all other parameters unchanged), the results will be different.
- You can set the Stochastic Seed as an integer in the Must-Have Properties of the General Simulation Properties.
- You should just try running the same model with a different stochastic seed and then compare the reports.
- It is indispensable to run multiple simulations on a model, each with different general seeds.
Local Stochastic Seed
- You can also assign an individual stochastic seed for every stochastic distribution.
- The individual seed is independent of the general seed.
- For example, giving the distribution of any activity an individual seed will result in the same values taken from this exact distribution in every simulation
- even though the general seed could have changed.