Further results of Gravitational Swarm Intelligence for Graph Coloring

Abstract

We provide enhanced results of an innovative nature inspired algorithm to solve the Graph Coloring Problem (GCP): the Gravitational Swarm Intelligence (GSI). Swarm Intelligence solves complex problems by extracting information from the spatial configurations of of agents that decide their actions/motions based on local available information, without any central control system. In the GSI application to GCP, agents correspond to graph's nodes, moving as particles in the gravitatory field defined by some target objects corresponding to graph node colors. Knowledge of the graph's topology is available to the agents' local navigation control as friend-orfoe information. We discuss the convergence of the algorithm and test it over well-known benchmarking graphs, achieving good results in a reasonable time.

Instances of subconscious social intelligent computing

“…every night. For clustering, a previously developed graph coloring algorithm based in Swarm Intelligence [l3], [14] is used. In figure 2 we can see the results on the a bigger ISN with 400 users.…”