Natural Selection Simulator using Machine Learning
Ramandeep Singh, Naveen Mehra, Akanksha Dhamija · 2022 Fifth International Conference on Computational Intelligence and Communication Technologies (CCICT) · 2022
Neural Networks and Machine Learning can be used in areas such as video games and simulators. In video games, the behavior of enemy AI can be made less repetitive by using machine learning. In this paper, using Unity Game Engine, a Natural Selection Simulator is created. The agent is given a neural network and acts as an organism whose aim is to collect a food pill that spawns on a platform (a simple plane). The agent themselves are spawned on the edges of the platform, while the food pills are spawned in the middle. A variation of the NEAT Algorithm is used to train the agent’s movement first. The movement includes leaving the edge of the platform, collecting the food pill, and going back to the edge. Once the movement has been trained, a simple genetic algorithm is used to evolve the agent’s speed in order to mimic how the traits of an organism evolve in nature based on its environment.