Wildfire Evacuation Routing (2D)
A wind-driven fire cellular automaton spreads across a road-grid network while evacuee vehicles are routed live to the nearest safe exit by a multi-source Dijkstra solver under BPR-style traffic congestion — watch roads burn shut and routes reroute in real time.
This 2D companion builds a real road-network evacuation model rather than reproducing the 3D original's decorative particle field: a 9×9 grid of intersections forms the road network, a wind-biased cellular automaton ignites and spreads fire across nodes (blocking every road that touches a burning or burned intersection), and a multi-source Dijkstra solver recomputes, every fraction of a second, the fastest route from every intersection to its nearest open exit under a BPR-style congestion function that slows vehicles down as traffic on a road approaches its capacity. Evacuee vehicles spawn continuously, follow the live routing table toward the grid's edge, and are marked evacuated, en route, or trapped depending on whether fire or gridlock cuts off every path out — so turning up wind strength or fire spread visibly forces traffic onto fewer, more congested roads exactly as the underlying network-optimization math predicts.
Wind-driven fire cellular automaton over a 9×9 road-grid graph; a multi-source Dijkstra router recomputes shortest safe-exit routes as roads burn out, under a BPR-style traffic congestion function.
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install