Evolving a Neural Network in a sparse reward environment

Evolving a Neural Network in a Sparse Reward Environment

Using genetic algorithms to solve the Lunar Lander Continuous environment with a sparse reward

Genetic algorithms are a powerful method, inspired by the biological evolution, to solve optimization problems. They consist of a sequential process creating solutions through random mutations and crossover and selecting the best solutions (according to a quantity to be maximized) for reproduction. A key element in the selection of the best …

environment genetic-algorithm machine learning network neural network neural networks reinforcement learning thoughts-and-theory