What is BoxCar 2D?
The program utilizes a genetic algorithm to learn and iteratively improve the design of a virtual car. It begins with a population of 20 randomly generated car shapes, each equipped with wheels. The program evaluates the performance of each car by running a simulation to measure how far it can travel. The cars that demonstrate the greatest travel distance are selected to "reproduce," with their design traits combined to generate offspring for the next generation. This evolutionary process continues, with the program progressively refining the car designs to enhance their performance
Highlights
- Genetic algorithm-based car design optimization
- Evaluation of car performance through simulation
- Selective reproduction to generate improved car designs over generations
- Ability to input custom car designs and test on different terrains
Platforms
- Online
Languages
- English