Skip to main content
Fig. 5 | Applied Network Science

Fig. 5

From: Uncovering the social interaction network in swarm intelligence algorithms

Fig. 5

The interaction network provides us the means to examine swarm-based algorithms from a general perspective. Here we use simple definitions of the interaction network for four different algorithms: the Particle Swarm Optimization (PSO), the Fish School Search (FSS), the Artificial Bee Colony (ABC), and the Ant Colony Optimization (ACO). Though they have different bio-inspirations (i.e., bees, ants, birds, and fishes), we can analyze them in the same interaction space. For this, we build the interaction network I(t) for each of them based on the social operators in the algorithm. We sum up each matrix over time to analyze \(\mathbf {I} = \sum _{t} \mathbf {I}(t)\). (a) After 700 iterations, each algorithm shows distinct signatures. (b) The strength of a node (i.e., \(\sum _{j} \mathbf {I}_{ij}\)) tells us the influence of an individual on the population. Though PSO and FSS allow strong influencers to exist, ABC and ACO exhibit a well-behaved distribution of spreaders

Back to article page