Closed FelipeRosa closed 8 years ago
Seems ineffective, programs still seem to grow too large.
Review and test more?
I think another method of limiting program size is still needed. Parsimony pressure seems to increase the quality of the solutions the algorithm finds, but program size still unbounded.
Maybe implement some hard limiting technique?
Worlds should be able to pass a covariant parsimony pressure coefficient to the selection function (while computing the next generation), so that program size can be limited.