一种基于等级有序竞争的群集控制

Flocking based on coordination of ranked agents

就异构群集内部的竞争协调问题,根据非移动状态下金丝猴群空间结构和等级关系的研究成果,构建了一种等级有序竞争智能体模型,并提出一种基于等级有序竞争的群集控制算法.算法通过利用质心Voronoi划分,将给定区域内等级个体的群集控制问题描述为一个关于个体位置的有约束优化问题;个体综合考虑自身安全、效用函数优化以及群集内等级有序协调3方面要求,通过模仿动物个体间基于等级的协调行为,有效地解决了群集过程中个体安全和异构个体间的冲突;最后获得的群集分布呈现等级有序性.理论分析和仿真实验结果验证了算法的有效性.

A model of competitive agent was proposed for the coordination of heterogenous agents during flocking inspired by the work on the relationships between the spatial structure and dominance of the golden snub-nosed monkey group in no-locomation. Then a new distributed coordination algorithm of ranked agents was presented. Flocking of ranked agents on a given region was formalized as a constrained locational optimization problem of each agent by using centroidal Voronoi partitions. Each agent integrated the objectives of safety, utility function optimization and hierarchized coordination. Then the individual safety was ensured and the conflicts among individuals during flocking were effectively resolved. The final flock was stable and hierarchized. The theoretical and experimental results proved the effectiveness of the algorithm.