So in my objective function, results are saved to the drive on every iteration. Right now I am using the tempname() function to create non-conflicting filenames. Because I am using parallel optimization, there is no easy way to match the files to the iteration number and thus to the results.
In the single threaded case I could just use a global variable or even just sort the files by date and time. But in parallel, it is unclear which thread is excuting which part first and this could lead to the file creation/modification times not matching the order of iterations.
The easiest solution would be to base the filename on the iteration number, but I haven't found a way of accessing this information in the objective function.
