Collective Intelligence and Other Extensions of Evolutionary Computation
David B. Fogel, Derong Liu, James M. Keller · 2016
This chapter covers the aspects of evolutionary algorithms and methods that are related to simulating evolution on computers. It begins with a population-based optimization approach called particle swarm optimization (PSO), which models the flocking behavior in certain animals. The chapter then focuses on another population-based approach called differential evolution, which searches a landscape for optima by using different vectors between existing solutions. The key ingredient in differential evolution is that individuals move based on the differential vectors from the individual to other individuals in the population. Another biologically inspired method for solving problems is ant colony optimization (ACO), and it simulates how ants discover food sources and communicate their discoveries with other ants. Finally, the chapter involves finding solutions that satisfy multiple criteria. One approach to handling multiple criteria is to combine them in a single utility function that returns a real value.