pyevolution 0.1.0

PyTorch based framework for training artificial agents in bio-inspired environments

PyEvolution

PyEvolution is a deep reinforcement learning framework for evolution simulation environments. PyEvolution provides a high level API for training agents using standard Deep Reinforcement Learning Algorithms in environments with in-built neural networks for performing policy evaluation. Users can thus focus on the main problem - to observe interesting patterns amongst these artificial species in various simulated environments.

Currently we have a single environment consisting of a single species called "Prima Vita" as a proof of concept. The details of this environment can be found below.

Prima Vita

Prima Vita is a species of artificially simulated beings created as part of Project DC. This repository holds a simulation environment created in pygame which is to be used with Deep Reinforcement Learning algorithms to find out the evolution of Prima Vita.