An implementation of NEAT based on Ken Stanley's implementation from 2002. This project is designed to teach itself how to play Snake and use Python for data analytics.
Created by Sheldon Woodward and Charles Lambert.
- Import topology dump
- Recurrence
- Multi-threading
- True random
- Percent population replacement
- More parent and survivor selection algorithms
- Trainer abstraction
- ANN copy constructor
- Implement vectors with std::unique_ptr instead of deques
- Make ConnectionGene a struct
- Put GA parameters into a struct
- Tests