Leader Follower Problem

We try to make the problem harder for WL and add the use of macrolearning.
In this experiment macrolearning consists of changing the subworld memberships of the agents. Each subworld has three agents.
In this problem each leader has two followers. The two followers should pick the same night as the leader.
Macrolearning is used after 500 weeks. At that time correlations are found between agents attendance and the most correlated are placed in the same subworld.