SWARM VS. MAZE / LIVE PATHFINDING

CHANGE THE WALLS.
WATCH THE REROUTE.

Hundreds of agents. One exit.
Reshape the maze while they move.

ENTRY / LEFT EDGELAYOUT 01EXIT / BOTTOM RIGHT

ESCAPED0 / 256

ELAPSED0.0s

FIRST FINISH—

MEAN FINISH—

AGENTSFOLLOW AGENT SHOW ITS PLANNED ROUTE CHANGE WALLS EVERY 30s

Agents start along the left edge. Each follows a shortest route to the exit.

Green: moving agents. Orange: the selected route. Completion times use simulation time; pause freezes the clock.

This is a separate algorithmic pathfinding experiment. Agents have full map knowledge and use shortest paths, not learned navigation or the homepage’s foraging rules. Changing the maze replans unfinished agents from their last cell center and clears their old trails. Finished agents retain their times; results from changed layouts are not a fixed-map benchmark.

CHANGE THE WALLS.WATCH THE REROUTE.

CHANGE THE WALLS.
WATCH THE REROUTE.