Reproducing migration app with foreach

In the previous chapters, you parallelized an application of migration projections. Now let's make sure the application is reproducible in both parallel as well as sequential environments. You will use the foreach framework with the doParallel, doSEQ and doRNG backends for the parallel, sequential and RNG environments, respectively. The function to use is ar1_block_of_trajectories() which you will instruct to generate 10 migration trajectories for each of the first 5 rows of the ar1est parameter dataset.

All the necessary functions, the dataset, seed (set to 123) and packages doParallel and doRNG are loaded in your workspace.

For the parallel run, register doParallel with two workers, as well as doRNG using the seed object.
Construct a foreach-%dopar% loop (with iterator r = 1:5) that runs ar1_block_of_trajectories(), passing r as its argument.
Register the sequential backend and set the seed to seed using set.seed() and run the same foreach loop but with %dorng%.
Check if results from the two steps above are identical.

Take Hint (-30 XP)

Exercise

Reproducing migration app with foreach

Instructions