As we all currently know, COVID vaccines are being dispersed throughout the world at varying rates and times. As a general area of interest, we wanted to know if the solution to this could lie in quantum computing. Could Qiskit Aqua, given a certain number of trucks and stops, propose ideal routes for delivery in a realistic way? Furthermore, if it became a traveling drone problem instead of trucks, could the delivery be done more efficiently?
What it does
Our API makes use of Google Maps to calculate distances and routes from one depot to drop-off locations. Using Qiskit Aqua, it gives a comparison of methods (the noisy QASM Simulator, the QASM Simulator with Matrix-State Method, the StateVector Simulator Method, and the classical/binary solver methods) for efficiency.
How we built it
Qiskit Aqua has a number of useful tutorials as great starting points. The one we used most was: https://qiskit.org/documentation/tutorials/optimization/7_examples_vehicle_routing.html
Challenges we ran into
Understanding the concepts and math required to apply the Vehicle Routing Problem to our setting, as well as making sure that the frontend and backend were communicating properly. We faced issues with Socket when we deployed the application, later on, we fixed the issue the demo video does not have the working example but you can try it out on the website now :)
Accomplishments that we're proud of
We are proud of learning how to use Qiskit Aqua and understanding its versatility, as well as creating something with an easy interface for users to be able to access this service.
What we learned
Classical adaptations of quantum methods are much slower than ideal classical solutions. Shocking.
What's next for Delivering Vaccines Using Qiskit and Vehicle Routing Problem
For the next steps, the main concern would be adapting this code into more realistic settings. For example, could Google Maps tell it the best route knowing that there might be road blockages or bad weather ahead?