Folded: Genetic Glider Simulation

Folded is a genetic algorithm based simulation that optimizes for distance travelled by paper gliders. I hope this project can be used to refine understanding of aerodynamic laws by synchronously running an agent that folds and throws paper gliders simulated in Folded to determine inconsistencies and highlight misconceptions with known physics laws.

Check out Folded here

Components of Project

Physics Engine

Folded relies on accurate calculations for aerodynamics experienced by paper gliders. This simulation assumes the environment exhibits laminar flow.

Genetic Algorithm

With the objective of optimizing for distance travelled away from starting point, a genetic algorithm with the fitness score being distance travelled is a perfect use case. Genes are mapped to various characteristics for each piece of paper.

Folding Algorithm

Paper characteristics are passed through a folding algorithm to get accurate glider objects. Triangle meshes are dynamically generated for obtainable paper configurations. The next iteration of this project would see rectangular wings, requiring more math complexity.

List of Managers

A camera manager, sound manager, game manager, and many others are used to manage and display information that is pertinent for users.

How did this project start?

I enjoyed physics from high school and previous college courses and saw a use case for the genetic algorithm in examples like genetic cars in my artificial intelligence class where the evolution is specific to the programmed environment. I wondered why people don't use this for more real life applications by making the environment as we understand real life to catch misconceptions and improve understanding of physics.

Mean Green Machines

Gliders are one of the most eco-friendly aircraft, as they don’t use fuel and rely entirely on nature for flight.