It should be possible to run simulations concurrently from R

[abdelr]

SimulationRunner should accept a list of simulations and run them concurrently.
SimulationBatch should accept a list of SimulationBatchRunValues and run them concurrently

[abdelr]

Changes to the original implementation:

1. Adding ConcurrencyManager class
   • This class should handle all concurrency in a way that you could specify what code to run over which data list on how many cores.
   • It should never create more threads than cores.
   • It should never create more threads than data elements.
   • Every thread should work for as long as there is still data to consume. The data is consume from a ConcurrentList. When a thread finishes processing a piece of data, it will try to take the new piece of data so if one thread needs longer to execute the other threads should take over instead of bein idle.
   • When the class is stable, PopulationRunner (or any further class in need of concurrency) should not handle the concurrency on its own but rather use this class.
2. Adding ConcurrentSimulationRunner class
   • The class should allow for setting the amount of cores to use (0 or negative means use maximum possible).
   • It should also allow for adding a Simulation object.
   • It should allow for adding a SimulationBatch object.
   • It should allow to run concurrently all added tasks (running a Simulation or SimulationBatch).