Inspiration

During emergencies such as floods, wildfires, and chemical leaks, sending everyone through the shortest route at the same time can create severe traffic congestion. We wanted to build a system that makes evacuation faster, safer, and more organized.

What it does

EvacuSafe creates an evacuation plan by considering road capacity, travel time, population, and shelter capacity. It decides when different groups should leave, which routes they should use, and which shelters they should reach.

The system also supports staged departures, so everyone does not enter the roads at once. This helps reduce congestion and overall evacuation time.

How we built it

We modeled the city as a graph where:

  • Nodes represent locations or zones.
  • Roads represent graph edges.
  • Each road has a travel time and capacity.
  • Zones contain populations.
  • Shelters have limited capacities.

We use a time-expanded flow approach to plan evacuation over time. The generated plan is then tested using a simple queue-based traffic simulator.

We also implemented dynamic replanning. If a road becomes blocked during evacuation, the system starts from the current locations of evacuees and calculates new routes instead of restarting the entire evacuation.

Challenges

The main challenge was balancing travel time with road congestion. The shortest route is not always the fastest when many people use it simultaneously.

Another challenge was respecting both road capacities and shelter capacities while scheduling departures.

Handling a road closure during an ongoing evacuation was also important. The planner must react to the new situation without losing track of people who are already moving.

What we learned

We learned how graph algorithms, flow optimization, scheduling, and traffic simulation can work together to solve a real-world emergency planning problem.

Our experiments compare EvacuSafe with an "everyone leaves immediately using the shortest route" baseline and measure total clearance time and average evacuation time.

Built With

  • algorithms
  • data
  • dynamic
  • evacuation
  • expanded
  • flow
  • graph
  • network
  • optimization
  • path
  • planning
  • python
  • routing
  • shortest
  • simulation
  • structures
  • time
  • traffic
Share this project:

Updates

Submission history