Analysis and Parameter Adjustment of the RDPSO Towards an Understanding of Robotic Network Dynamic Partitioning based on Darwin’s Theory

Corresponding author Member of the Institute of Systems and Robotics, University of Coimbra, Portugal Member of the RoboCorp, DEE, Engineering Institute of Coimbra, Portugal Member of the Knowledge Engineering and Decision Support Research Center, Porto Portugal. 1588 M. S. Couceiro, F. M. L. Martins, R. P. Rocha and N. M. Fonseca Ferreira Although the well-known Particle Swarm Optimization (PSO) algorithm has been first introduced more than a decade ago, there is a lack of methods to tune the algorithm parameters in order to improve its performance. An extension of the PSO to multi-robot foraging has been recently proposed and denoted as Robotic Darwinian PSO (RDPSO), wherein sociobiological mechanisms are used to enhance the ability to escape from local optima. This novel swarm algorithm benefits from using multiple smaller networks (one for each swarm), thus decreasing the number of nodes (i.e., robots) and the amount of information exchanged among robots belonging to the same sub-network. This article presents a formal analysis of RDPSO in order to better understand the relationship between the algorithm’s parameters and its convergence. Therefore, a stability analysis and parameter adjustment based on acceleration and deceleration states of the robots is performed. These parameters are evaluated in a population of physical mobile robots for different values of communication range. Experimental results show that, for the proposed mission and parameter tuning, the algorithm con-verges to the global optimum in approximately 90% of the experiments regardless on the number of robots and the communication range. Mathematics Subject Classification: 39A30, 70E60, 65L20