Maturity of the Particle Swarm as a Metric for Measuring the Particle Swarm Intelligence

Particle Swarm Optimization, Maturity Model, Particle Swarm Intelligence

In the research work titled Maturity of the Particle Swarm as a Metric for Measuring the Particle Swarm Intelligence, the cooperative Particle Swarm Optimization algorithms with recpect to a maturity model derived from the Maturity Model of Command and Control operational space and the model of Collaborating Software have been analyzed. Cooperative swarming strategies directly derived from the original version of the algorithm published in 1995 have been studied. It has been concluded that the particle swarm is not mature enough because of the lack of system's awarenes, and that a solution would be some adaptation of particle's behavioural rules so that the particle could use coefficients for updating its velocity whose value would be provided from inside the swarm system, without external control. The true intention of the research is to contribute to a deeper understanding, from which the swarm derives its intelligent behaviour.

The PSO algorithm is known primarily as a stochastic algorithm which shows signs of intelligent behaviour. In this paper, a maturity model of the particle swarm operational space as a metric for the swarm intelligence is introduced and then, the swarm intelligence is assessed according to this model. The model is proposed as a combination of the Maturity Model of the C2 (Command and Control) operational space and the model of Collaborating Software. The swarm particles are considered for a system's software modules, and the PSO algorithm for the system. Then, the whole particle swarm is assessed according to generic characteristics of the collaborative behaviour. A three-dimensional maturity vector determines the resulting maturity of the swarm operational space. The coordinates of the maturity vector determine the degrees of allocation of decision rights, information sharing, and social interaction among particles. The conclusion from the maturity assessment is that the swarm system as a whole is not mature enough because of lack of the system's awareness. In order to improve the system's awareness, the decision making autonomy of individual particles should be enhanced so that the swarm could be considered as coordinated or collaborative, eventually.