Post-earthquake routing is not shortest-path with traffic. Debris, aftershocks and incomplete information demand policies that stay good when the map is wrong.
Navigation apps solve routing on a network that is known and mostly intact. After a major earthquake, neither assumption holds. Streets are blocked by debris that no map shows yet, bridges await inspection, and an aftershock can invalidate a route that was clear twenty minutes ago.
The naive fix — recompute shortest paths as reports arrive — fails in a subtler way: it optimizes against a map you know is wrong. In our simulations, dispatch policies that chased the momentarily-optimal route often performed worse than policies that accepted longer but structurally safer corridors.
What works is robustness. Routes along wide arterials with redundant alternatives degrade gracefully; clever shortcuts through narrow streets fail catastrophically. Policies that keep ambulances on a maintained "lifeline subnetwork" — pre-identified, prioritized for clearance, continuously verified — beat globally optimal routing under almost every uncertainty scenario we tested.
The operational conclusion is that routing after an earthquake is a planning problem before it is an algorithm problem. Choose and harden your lifeline corridors now; the best dispatch policy on the day is the one that already knows where it is allowed to trust the map.
Comments
No comments yet — be the first.