Today I made a new prototype to explore and display the behaviour of the program that scans the PEAC statespace from many starting points. Have a look at http://ygdes.com/~whygee/a2_02.html
Here are a couple of videos to show the behaviour when the number of parallel jobs is modified. In the worst case, with only two jobs, it takes a lot of time (all the steps to scan the whole orbit).
When more jobs are possible, it becomes interesting... Here, each colour represents a job and not an orbit.
Some arcs are extremely short, others -rare- are very long, so they waste most of the time.