基于有限记忆的智能群体群集运动控制  

Control of Flocking Motion of the Intelligent Swarm with Limited-memory

作　　者：王冬梅[1] 方华京[1] 

机构地区：[1]华中科技大学控制科学与工程系,武汉430074

出　　处：《武汉理工大学学报（交通科学与工程版）》2010年第1期122-125,共4页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国家自然科学基金项目资助(批准号:60574088)

摘　　要：以历史状态信息来模拟刻画个体的个性特征,将组成群体的简单个体扩展为具有有限记忆能力的智能个体,研究了个体的个性特征对群集行为的影响;建立了具有简单智能的群体的群集控制模型并进行了稳定性分析.智能个体根据记忆单元的历史状态信息利用最小二乘法预估计个体的下一步运行状态,并以此作为虚拟领航者引导群体运行,实现智能群体群集运动控制.仿真结果验证了该模型的有效性,证实了个体的个性特征使得群体最终在稳态附近波动.Using historical states to mimic individual character of agent and giving the individual agent a rudimentary intelligence in the form of a limited memory, the influence of individual character on flocking motion is studied, and an extension to a flocking control model is built, and stability analysis is given. According to the past states in the memory space, least square method is used to predict the fu- ture state, and the future state is a virtual leader leading the agent to achieve flocking motion. The results of simulations have shown that the algorithm is valid, and have testified that individual character of agent makes the swarm fluctuate at the state of stability.

关 键 词：群集运动 最小二乘法 虚拟领航者 稳定性 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]