I am implementing an NSGA-II using Platypus. I am not able to show all my code, but I have close to 30 decision variables, 2 objectives, 1 constraint.
It would run smoothly for the first 200 evaluations of the function, but then returns the error:
TypeError: unhashable type: 'numpy.ndarray'
I reckon it is because the way NSGA-II is implemented by Platypus, as each generation consists of 100 population and 200 corresponds to 2 generations of iteration. But what reasons would make this error appear at the 3rd generation?
Think I found the reason myself. It was not because of the Platypus implementation. Instead, I made a mistake by introducing a 'float+list' type objective