qhgordon / apply-ai-technique-to-airline-scheduling

Apply-AI-technique-to-Airline-Scheduling

Proplem Setting

With the mega storm last Christmas, there was a cascade of flights cancelled -- thousands! How would you figure out what the planes and flight crews should do, given the uncertain nature of air travel?

Constraint Programming

The variables in this scenario could be the flight schedules, the flight crews, and the planes.

The domains of the variables could be the possible flight schedules, the flight crews available, and the planes available.

The constraints could be the availability of flight crews, the availability of planes, and the minimum rest time required for flight crews.

Bayes Nets

The random variables in this scenario could be the weather conditions, the number of flight cancellations, and the flight schedule changes.

The variables could be connected such that the weather conditions affect the number of flight cancellations, and the number of flight cancellations affects the flight schedule changes.

The hidden variables could be the weather conditions, and the observed variables could be the number of flight cancellations and the flight schedule changes.

The Bayes Net could be used to make predictions about the number of flight cancellations and the flight schedule changes given the current weather conditions.

Deep Learning

The architecture that could be used in this scenario could be a neural network that takes in the weather conditions as inputs and outputs the number of flight cancellations and the flight schedule changes.

The training data could be collected from previous flight cancellations and flight schedule changes and used to train the neural network.