Goal Oriented Action Planning

Features

Templated world values

I needed a light weight Variant type so I could store a list of different typed values in my WorldState, for this I used inheritance with a interface and templating. Source File

Cyclic graph in planning stage

When planning the GOAP planner looks at every action that can be performed on the current world state and creates a graph node for that resulting world state, then adds an edge with the action performed.
This process is repeated on every resulting world state and for any reoccurring world states the a node is not added but the existing node with that world state is connected instead.

Dijkstra's to find the best plan

Once the graph is created I use Dijkstra's to search and find the path of actions that results in a goal meeting world state.

Source Code