Many works and my experience in my last work show that people use some kind of hierarchical approach in their learning and decision making process. As it has the both benefits of flexibility due to environment changes and the lower computational cost. So I decided to work in this field for my thesis.
We designed a novel task and analyzed the human behaviors and their eye tracking data while performing the task. We saw that in some complex situations people would define some irrational subgoals to reduce the complexity while there is no distinct subgoal defined for them.