PERCEPTS ⇄ ACTIONSthe world: sensed under the Observability setting — and the path the agent finds is walked right here
Escaped!
hover the maze to inspect cells
slow
fast
AGENT FUNCTIONAgent
heuristic is used by Greedy · A* · Hill climb (falls back to Manhattan on errors)
Metrics
| steps | 0 |
| path cost | 0 |
| optimal steps | — |
| optimal cost | — |
| nodes expanded | — |
| frontier size | — |
| score | — |
ENVIRONMENT
Observability
Determinism (M3)
Environment scorecard (live)
| observable | Fully |
| deterministic | Deterministic |
| agents | Single agent |
| episodic / sequential | Sequential |
| static / dynamic | Static |
| discrete / continuous | Discrete |
every row is a design decision — flip one, and a different algorithm family applies. The agents row flips in Game Arena.
RENDERING
Theme
Walls
View