Details
Contribution
This tech demo was created to study the Sense-Decide-Act Loop which is used for managing large quantities of AI agents. A simple game of life was created for testing the implementation as well. This pattern allows for dividing the update of an AI agent into 3 different stages that each can be ticked on a different frame. Sense stage serving as the reaction of an agent, Decide as the thinking capabilities of the agent and Act being the action speed of the agent. Each of these stages can be ticked at different intervals for greater control and performance. The pattern also allows for ticking a specific amount of agents instead of all active agents each frame which helps out tremendously performance wise since it avoids spikes in frame time on the frame when all agents update. Lastly, flocking behavior was implemented for the birds which is a simple behavior that can be used for a lot of different AI agents.