Guides

AI & simulation tools

Behavior trees, MCTS, resource allocation, FABRIK IK, and spatial queries.

This page is a stub.

Meep ships a set of decision-making and simulation tools that are rarely found in JavaScript engines:

  • Behavior trees and blackboards — industry-standard state management for AI agents.
  • Monte-Carlo Tree Search — the same family of algorithm AlphaGo used. Suitable for turn-based AI, planning, and game theory problems.
  • Resource allocation solver — plan optimal action sequences under resource constraints (ammo, currency, health, time).
  • FABRIK inverse kinematics — fast, robust IK with bipedal and quadrupedal specializations for terrain adaptation.
  • BVH-based spatial queries — ray, box, frustum, and point queries against the scene at any scale.

Most of these live under src/engine/intelligence/ in the source tree — behavior trees, MCTS, and the resource allocation solver each have their own subdirectory with self-contained tests. The FABRIK solver lives separately under src/engine/physics/inverse_kinematics/fabrik/, with higher-level IK component wrappers (two-bone, surface alignment) under src/engine/ecs/ik/.