Can Complex Collective Behaviour Be Generated Through Randomness, Memory and a Pinch of Luck?

Abstract

Machine Learning techniques have been used to teach computer programs how to play games as complicated as Chess and Go. These were achieved using powerful tools such as Neural Networks and Parallel Computing on Supercomputers. In this paper, we define a model of populational growth and evolution based on the idea of Reinforcement Learning, but using only the 3 sources stated in the title processed on a low-tier laptop. The model correctly predicts the development of a population around food sources and their migration in search of a new one when the known ones become saturated. Additionally, we compared our model to a pure random one and the population number was fitted to a logistic function for two interesting evolutions of the system.