At the lowest level (which is what is currently implemented), the simulation should simply store simulation results in the specified directory. As implemented, the saving to a subdirectory not what I would expected to happen. If desired, you could check that this directory is empty so that previous aren't overwritten.
Higher level functionality for running simulation batches could create multiple subdirectories named with their index within the sequence of simulations (or timestamp, but I think index is more informative than timestamp)
At the lowest level (which is what is currently implemented), the simulation should simply store simulation results in the specified directory. As implemented, the saving to a subdirectory not what I would expected to happen. If desired, you could check that this directory is empty so that previous aren't overwritten.
Higher level functionality for running simulation batches could create multiple subdirectories named with their index within the sequence of simulations (or timestamp, but I think index is more informative than timestamp)