Rohit Punjabi

The tech demo consists the implementation of combining GOAP with Smart Objects for this demo we decided to have the planner compile a list of all Smart Objects in the scene, and assign preconditions and effects to Smart Objects directly, instead of defining a separate “Action” class. Then, when agents look for a plan they query the Planner with their current state and their goal state. The Planner then generates a list of the Smart Objects the agent should interact with and the resulting states between interactions. This list is ordered so that each interaction moves the agent’s state towards the goal state in a believable manner. The agent then utilizes an internal Finite State Machine (FSM) to iterate through the list, interacting with the Smart Objects in order. Our implementation was done using Unreal Engine 4, although it could be applied to almost any C++-based engine/editor.

Video of Demo

Video of Explanation