A complex neighborhood based particle swarm optimization
CEC'09 Proceedings of the Eleventh conference on Congress on Evolutionary Computation
Scale-free fully informed particle swarm optimization algorithm
Information Sciences: an International Journal
The fully informed particle swarm: simpler, maybe better
IEEE Transactions on Evolutionary Computation
| Hi-index | 0.00 |
In this work we analyzed the role social networks play in the efficiency of collective problem-solving, evaluating whether the topological characteristics seen in real-world networks yield any performance improvement in such processes. To study this we used the Particle Swarm Optimization as a testbed for social groups performing a collective task, defining the structure of communication between individuals in the swarm through topologies generated by a model for the creation and evolution of social networks. The experimental results indicate that groups using these networks may, indeed, experience better performance in collective problem-solving, so that these groups were able to overcome the results achieved by swarms using classical neighborhoods for PSO and reached results very close to those found by swarms using the topology of DMS-PSO, usually considered to be part of the state-of-the-art of Particle Swarm Optimization.