instantiate stan model with same random seed as used for sampling

[mitzimorris]

Summary:

Please provide a short couple sentence summary.
In order to use RNG functions in the tranformed data block, the Stan model must be constructed with a boost rng object which has been instantiated with the random seed.

Description:

see stan-dev/stan#2241 for details and discussion.

Reproducible Steps:

Use the following program to test:

data {
  int<lower=0> N;
}
transformed data {
  vector[N] y;
  for (n in 1:N)
    y[n] = normal_rng(0, 1);
  print(y);
}
parameters {
  real mu;
  real<lower = 0> sigma;
}
model {
  y ~ normal(mu, sigma);
}

Current Output:

Current output results in identical values for vector y, no matter what the seed.

Expected Output:

Different seeds should give different values.

Current Version:

v2.15.0