joneshshrestha/Evolution-Simulator

The simulator will be built with Javascript as primary development language aided by HTML, CSS and other libraries for frontend purpose. As our Machine learning library we’ll be using Tensorflow.js, a low level ML library recently acquired by Google Inc. Our main objective is to build a graphics environment that enable users to create a completely artificial creature that will evolve to walk, run or even jump over obstacles to reach an objective point. Users will be given tools to create Nodes, Bones and Muscles which will help them generate an artificial creature. Just as in the process of Natural Selection of Evolution, multiple creatures are created with the same structure but different brain. These batches of creatures are called generation. All these creatures will try their best to achieve maximum speed to reach the objective point. Through the process of crossover and mutation, few creatures that performed better than the rest will be chosen and using their genes new offsprings are created. This process is continued until certain error rate or fitness level is reached.