I have an idea for optimize() to use multiple CPU cores.
For every test parameter set, you can open a new Zorro process to test and get results back.
And you can limit the process numbers, Just like WFO using NumCores
Could you do that?

That is theoretically possible, but less effective than the current WFO based multicore support. Creating a process is slow, while a single parameter run is normally fast. So doing it in a separate process will not save much time.