Mixed-integer linear programming model to schedule flights and distribute aircraft to minimize cost.

Run instance_gen.py to generate an instance.
You can edit the specific values to your liking in the CSV files under src/instance. Run lp.py to find a solution.

You can alternatively call the lp.solve_instance function directly by passing in your instance data (aircraft, airports, and routes).

Decision variables are $x_{i,j}$ where $i$ is the aircraft index and $j$ is the route index and $\forall x_{i,j} \in \left\{ 0,1 \right\} $

Let $a_i$ be aircraft with index $i \in \mathbb{Z}^+$ and let $r_j$ be route with index $j \in \mathbb{Z}^+$.
I approximate the variable cost, $c$, of running a route as $$c_{i,j}=x_{i,j} \cdot \text{efficiency} \left( a_i \right) \cdot \text{distance}\left(r_j\right)$$

The objective function is $MIN_{i,j} (c)$

The model considers the following constraints:

• each route must be covered exactly once
• the solution cannot use more aircraft than described in the fleet
• aircraft can carry strictly less than their maximum payload (cargo + passenger)
• aircraft cannot service a route who's distance is larger than the aircraft's range

All airports and aircraft in the model are real-to-life. Distances are calculated and cached at runtime based on airport locations using the haversine formula. Routes are strictly unique and route A → B $\ne$ B → A.

• "Our Airports" Open Dataset
• Various manufacturer resources for aircraft specs

• route visualizer
• move to pax fuel efficiency
• nonlinear fuel function
• rotation problem
• aircraft maintenance

Aly Ashour