论文部分内容阅读
针对现有的多智能体网络牵制蜂拥算法中网络在演化过程中可能出现“分裂”的问题,以代数连通度作为衡量网络连通的指标,提出一种基于局部优化的牵制蜂拥控制算法,以保持多智能体网络在演化过程中的全局连通性,从而保证所有智能体跟随虚拟领导者运动.该算法将局部优化策略引入牵制蜂拥控制器中,考虑网络中仅有部分智能体(称为信息智能体)具有虚拟领导者的信息,使信息智能体不断朝着其在当前时刻和下一时刻所有邻居的并集所组成网络的代数连通度最大的方向运动,以通过优化局部连通性达到提高整个网络连通性的目的.通过理论分析和仿真实验并与现有牵制蜂拥控制算法相比,进一步证明了算法可有效保持多智能体网络在演化过程中的全局连通性,且所有智能体均聚集在虚拟领导者周围更小的范围内,更能体现信息智能体在牵制蜂拥控制网络中的相对重要性.
In order to solve the problem of “split” existing in the existing multi-agent network embedding algorithm, the algebraic connectivity is taken as an index to measure network connectivity. A constrained flocking control algorithm based on local optimization is proposed. In order to keep the global connectivity of multi-agent networks in evolution so as to ensure that all agents follow the movement of virtual leader.The algorithm introduces the local optimization strategy into the dene controller, and considers only part of the agents in the network Information Agent) has the information of the virtual leader, so that the Information Agent continuously moves toward the algebraic connectivity of the network formed by the union of all the neighbors of the current moment and the next moment, so as to reach by optimizing the local connectivity Improve the overall network connectivity.According to the theoretical analysis and simulation experiments and compared with the existing control algorithm of containment flocking, it is further proved that the algorithm can effectively maintain the global connectivity of multi-agent networks in the evolution process, and all the agents Gathered in a smaller area around the virtual leader, but also better reflect the information agent in the control network fob Relative importance.