There is currently one test checking if best_y is non-decreasing in subsequent generations. It should be verified if with current setup optimization really progresses. Adding tests for that may be hard as it needs to be shown that progress is eventually made. Additionally it may be worth showing that certain setups outperform their random counterparts.
There is currently one test checking if best_y is non-decreasing in subsequent generations. It should be verified if with current setup optimization really progresses. Adding tests for that may be hard as it needs to be shown that progress is eventually made. Additionally it may be worth showing that certain setups outperform their random counterparts.