Hopefully Fixed Issue with Experiment Execution

We wanted to use "spawn" as multiprocessing context, because "fork" is deprecated.
However, "spawn" cannot be used as it cannot pickle lambdas and almost all of our experiments use lambdas to generate algorithms and instances.
Without changing almost all code for experiment generation, right now, "spawn" is not usable.
In the long-term future, we will have to change the experiment generation facility entirely, probably to launch the python interpreter with the same command line arguments as the original process but with using only 1 CPU core or something.
This will have several other implications that need to be dealt with.
But it cannot be done now.