Multi-agent planning using HTN and GOAP
Abstract: This thesis examines how Hierarchical Task Networks (HTNs) can be used to
plan for overall strategies in a game environment where the agents are
controlled by GOAP (Goal Oriented Action Planning) at an individual level.
To solve this, a layer between a HTN planner and a game engine has been
implemented, the Planner Manager. A simple HTN planner has also been
modified to be able to handle parallel tasks. The results suggest that HTNs
can be used, it will decrease the work load for the programmers but increase
it for the designers during the production process. This is good because the
designers should be the ones that are in control during the production phase.
CLICK HERE TO DOWNLOAD THE WHOLE ESSAY. (in PDF format)