Update random number generator seeding in runsim by seeding a list of default_rng objects that are passed through to parameters that need to be sampled within each individual makesimpars to both ensure consistency and to avoid all parameters being sampled based on the same seed (e.g. all low or all high)
When sampling, each parameter will use its own generator seeded by the global seed + a hash of the parameter short name - this means that changing which parameters are sampled will still be consistent for other parameters for the same seed, resulting in more consistent outputs
runsimby seeding a list ofdefault_rngobjects that are passed through to parameters that need to be sampled within each individualmakesimparsto both ensure consistency and to avoid all parameters being sampled based on the same seed (e.g. all low or all high)