PSO is used to adjust the network weights, with the Adaptive Neural Swarming method, the controller could adapt to environmental changes. It is tested in a real-world task of controlling a simulated non-linear bioreactor.
A hybrid PSO is proposed by introduction of the natural selection mechanism. The authors demonstrated hybrid PSO is better than original version on distribution state estimation problems.
A paper by Frans van den Bergh and Andries P. Engelbrecht, South Africa. The interesting point is that they split the input vectors to several sub-vectors, each which is optimized cooperatively in its own swarm.
A concept for the optimization of nonlinear functions using particle swarm methodology by James Kennedy and Russell Eberhart. This is the first paper talking about PSO.