Online velocity optimization of robotic swarm flocking using particle swarm optimization (PSO) method

摘要

In this paper, the agent velocity in robotic swarm was determined by using particle swarm optimization (PSO) to maximize the robotic swarm coordination velocity. A swarm as supposed here is homogenous and includes at least two members. Motion and behavior of swarm members are mostly result of two different phenomena: interactive mutual forces and influence of the agent. Interactive mutual forces comprise both attraction and repulsion. To be more realistic the field of the swarm members' view is not infinity. So influence of the coordinator agent on the robotic swarm would be local. The objective here is to guide the robotic swarm with maximum possible velocity. According to equation motion of the system, this maximum value cannot be analytically obtained. PSO is a novel method in optimization which inspired from observations of birds flocking and fish schooling. As compared to other robust optimization methods, PSO is more efficient, and requires fewer number of function evaluations, while leading to better or the same quality of results. The results show the high ability of this evolutionary algorithm in solving complicated dynamic optimization problems.