Those problems are getting pretty big (e.g., 3000 species) when the base problem bound is being checked but limited to 2000 species.
- Just tried bumping the base max_allowed_num_spp down to 700 and the wrap came out around 2100 species, which is more reasonable.
May want to create a separate bound for each problem type.
Could also add a penalty term in the wrap generation EF so that it tries to keep the total number of species down.