jacobha/boids

Papers:

• https://esajournals.onlinelibrary.wiley.com/doi/epdf/10.1002/ecs2.2447
• https://ai.stanford.edu/~ang/papers/icml00-irl.pdf
• https://openreview.net/pdf?id=nosngu5XwY9

Code:

• https://github.com/qzed/irl-maxent/blob/master/

Action Space:

• create a 2D grid for discretizing the trajectories of each flight.
• implemented movement in all 8 neighboring cells.
• grid size: 8 (rows/latitude) x 7 (cols/longitude) = 56 total possible states. Each grid cell is 1000m x 1000m.
• action data is stored in ./action_space.csv. Format: Gen = generation of the bird chain, Pair = replicate, Release = flight number, Action = ordered states that a flight traverses, arranged from start to finish.
• Start grid ID = 54, End grid ID = 2