Structured particle swarm optimization and its application to control designs

Particle Swarm Optimization (PSO) is one of the population-based stochastic computational techniques, inspired by the bird flocking, fish schooling and swarming behavior. However, PSO is "non-structural" and, as far as the control design is concerned, it would be desirable to give the structure property to PSO. In this paper we propose a new type of population-based technique, which is given the structure of trees. That is why we call it "structured PSO". We also demonstrate by simulations that such a technique is applicable to standard control problems such as the Van der Pol problem, the singular control problem, and so on.