JPSPlus is a high-performance 2D grid pathfinding library for Roblox based on Jump Point Search Plus (JPS+) algorithm. It preprocesses static maps to enable very fast A* queries by “jumping” between critical nodes while preserving optimal paths on uniform-cost grids, with optional path expansion and smoothing for Dummy movement.
The Concept:
Traditional Jump Point Search (JPS) was originally developed as a superior competitor to standard A* for 2D grids, designed to speed up search by “jumping” over redundant nodes rather than checking every neighbor. JPS+ takes this evolution a step further by using preprocessing to pre-calculate these jump distances. This results in ultra-fast queries that maintain A*'s optimality but with significantly reduced runtime overhead on static maps.
Why did I create this? I noticed there werent any open-source “Jump Point Search” pathfinding algorithms yet for roblox, so I decided to bridge that gap.
# When to use
Best for:
- top-down / flat worlds
- maze/indoor navigation on a plane
- many repeated queries on mostly static maps
# When NOT to use
Avoid if:
- dynamic maps: If your map changes every few seconds (e.g., destructible terrain), the cost of repeatedly “baking” the map data outweighs the search speed benefits.
- Weighted Terrain: If you need movement penalties (e.g., “mud is slower than grass”), use standard A*. JPS relies on uniform costs to skip nodes safely.
- complex 3D verticality: This is strictly for 2D grids; it does not handle multi-floor navigation natively without logical separation.
# Limitations
- 2D grid navigation (X/Z plane); not a 3D navmesh
- Static or mostly-static obstacles (changing obstacles require rebuild or custom updates)
- Uniform movement costs (no weighted terrain without modification)
