Learning linear operators by coordinated aggregation-based PSO

The coordinated aggregation-based particle swarm optimisation (CA-PSO) technique is implemented for learning linear operators. In the developed learning algorithm, the swarm consists of linear operators instead of particles. Each candidate operator in the population is attracted only by operators with better objectives than its own; with the exception of the operator with the best achievement which moves randomly as a 'crazy' agent. The CA-PSO learning algorithm learns optimal quantum and other linear operators and its results are compared with those given by Dan Ventura's learning algorithm and genetic algorithms. The comparison proves an improved performance of the developed learning algorithm over the state-of-the-art methods.