Inspiration
I was inspired when I saw a twitter post detailing how someone was able to run a full ATC simulation on the ESP32-S3 microprocessor of a toothbrush, and it got me thinking about whether the problem would run well on a QPU.
What it does
It optimizes simplified runway landing schedules using classical preprocessing and QAOA, then validates the resulting schedules.
How we built it
I was trying to rush to get this built before the end of the Fall Fest so I used opencode with an AI backend of Kimi K3.1 on my local laptop to work with me to plan and execute this idea. It started with a messy noisy 7 qubit approach, but was able to clean up into a nicer clean 4 qubit approach that resulted in much less noise.
Challenges we ran into
The biggest challenge I ran into was time, since there's not a ton of time to learn quantum mechanics and code for a hackathon, but I was able to overcome it by using AI assistance. Another challenge I ran into was the incredibly low accuracy of the initial approach since the 7 qubit approach resulted in a lot of noise. The solution was to find a solution that used less qubits and therefore less noise.
Accomplishments that we're proud of
I'm proud of the ability to get 68% accuracy on a 1024 shot run on the first real quantum program I've made.
What we learned
I learned during the research phase that quantum brute forcing is not quite as strong as I had thought with it only being able to crack hashes in O(sqrt(n)) time using Grover's algorithm and is not an appreciable threat to encryption security in its current state. I also was able to see the effects of noise on my circuit in action as there was a large noise reduction when I scaled back my circuit from 7 qubits to 4 qubits. I definitely learned more about quantum computing as a whole and I know now that qiskit and IBM QPUs are waiting for the next big idea!
What's next for Quantum Optimization for ATC Runway Scheduling
The next step is likely to try and see if I can add more planes to the simulation and see how much farther I can go without getting impeded by noise.
Built With
- ai
- opencode
- python
- qiskit
