Application of interactive genetic algorithms to boid model based artificial fish schools

Abstract

In this paper, we present an extended boid model for simulating the aggregate moving of fish schools in a complex environment. The boids model is an example of an individual-based model. The global behavior of the school is simulated by a large number of interesting individual boid (fish). In our proposed model, each boid is an agent that following six behavior rules: avoiding collision against schoolmates; gathering together; following a feed; avoiding obstacle; avoiding enemy boids; boundaries. The moving vector of each boid is a linear combination of five behavior rule vectors, and the coefficients are optimized by using an interactive genetic algorithm (IGA). Unlike the classical GA, interactive GA can adopt a user's subjective evaluation as fitness, which is useful when a fitness function cannot be exactly determined. © 2008 Springer-Verlag Berlin Heidelberg.