Social interaction is a powerful optimiser: The particle swarm

The particle swarm algorithm helps computer scientists and engineers solve hard problems, and also suggests insights into the computational power of social interaction. The method evolved out of social psychological simulations of social influence processes featuring a positivity bias in self-presentation and a proclivity for mimicry, with some random noise added. In this talk I will introduce the algorithm, show its connections to social behaviours in humans and other species, and discuss effects of variations in rules and parameters. I will tend to emphasise research that simplifies the algorithm rather than making it harder, as researchers come to understand what is essential for intelligence to emerge in a population of simple interacting entities.