Ji-Rath/MassAIExample

What it solves

This project provides a collection of practical examples and research notes for implementing high-performance AI and entity management using the Mass ECS (Entity Component System) framework in Unreal Engine 5. It demonstrates how to handle thousands of entities simultaneously without sacrificing performance.

How it works

The project implements various simulation patterns using Mass processors, fragments (data), and traits. It leverages the State Tree for AI logic management, Smart Objects for environmental interaction, and Niagara for efficient visualization of up to 50,000 entities. It also includes a custom spatial hash grid for fast proximity queries and integrates RVO2-based collision avoidance.

Who it’s for

Game developers and technical artists working with Unreal Engine 5 who want to implement large-scale crowd simulations, complex agent behaviors, or high-performance entity systems.

Highlights

• Open Simulation: A demo where agents gather resources to build a house using smart objects and state trees.
• Massive Visualization: Efficient rendering of ~50k entities using Niagara, bone vertex animation, and billboards.
• Advanced Navigation: Integration of Nav Mesh with asynchronous functionality to reduce overhead.
• Spatial Optimization: A custom spatial hash grid for fast world queries and entity avoidance.
• State Tree Integration: Examples of organizing AI logic through tasks, nodes, and transitions.