复杂多自主体网络的多协调性与控制

Instead of singleness, multiplicity is usually the main character of the coordination in biological communities in nature. Compared to the single-coordination behavior, the collective dynamics of multi-coordination is richer and more complex, and is more valuable in the application of cooperation control of groups with multi-objectives or multi-tasks. In this project, base on the generating mechanisms, the evolution processes, and the final forms of multi-coordination in biological communities, the effects of network topology parameters and network-induced factors on the multi-coordination will be analyzed for the complex multi-agent networks. The relationships between the network topology parameters and multi-consensus will be revealed. By taking the network-induced delay as a parameter, the parameter boundaries for the multi-consensus will be derived, the division of the parameter space for different consensus will be analyzed and the bifurcation behaviors will be studied; The effects of network parameters and nodes' dynamics on the multi-coordination will be revealed for the complex multi-agent networks. The effects of network parameters and nodes' dynamics on the multi-synchronization will be analyzed, and the multi-synchronization phenomenon will be analyzed for complex multi-agent networks in which the agent nodes have multi-attractivity. The criteria of multi-synchronization will be established; The formation control strategies of multi-agent networks with multi-objectives will be proposed; The robust controller and the quantization controller will be designed for the multi-coordination of complex multi-agent networks with disterbances, uncertainties, communication constraints and other factors. This project is helpful in deeply understanding the existence of multi-coordination in biological communities. It has extended the contexts of the multiple collective behaviors of complex networks. It has important theoretical meaning and application values in the cooperation control and multi-objective-based decision making of multi-functional complex multi-agent networks.

自然界中生物群体的协调性往往并不单一而具有多样性。与单一的协调性相比，多协调性的群集动力学更加丰富和复杂，在多目标多任务下的群集协调控制中更有应用价值。本项目拟以生物多自主体为背景，基于生物群体多协调性的产生机制、演化过程和表现形态，研究网络拓扑参数及网络诱导因素影响下复杂多自主体网络的多协调性，建立网络初值及拓扑参数与多一致性间的对应关系，以网络诱导时滞为参数分析多一致性的边界条件，分析边界的分叉情况；研究节点复杂行为和网络共同作用下的多协调性，分析节点动力学和网络参数对产生多同步的影响，分析自主体节点具有多吸引性时的多同步现象，建立多同步判别准则；研究多目标编队控制策略；研究干扰、不确定、通信约束等因素影响下的多协调性鲁棒控制器和量化控制器的设计。本项目拓展了复杂网络群集协调性的多样性内涵，对解决多功能型复杂多自主体网络的协调控制和多目标决策支持具有重要的理论意义和实用价值。

复杂多自主体网络的协调性往往并不局限于某一种形式而具有多样性。与单一的协调性相比，多协调性的群集动力学更加丰富和复杂。已有关于复杂多自主体网络单一协调性的研究已取得了丰硕的研究成果，但对多协调性的研究还不够系统深入。本项目针对网络诱导因素和个体复杂行为影响下复杂多自主体网络的多协调性开展研究，包括多稳定性、多吸引子、多一致性、多同步和多目标编队。具体的，项目揭示了网络多重复杂连接方式下神经网络多稳定性的产生机理，建立了保守性较弱的多平衡态充分条件；提出了在多维空间上扩展实现多涡卷吸引子及多翅膀吸引子的系统化方法；分析了网络通信时滞、通信不连续性和随机因素影响下的多一致性，设计了若干分布式控制协议，分别实现了模一致性、二分一致性、任务驱动多一致性等；提出了初值依赖的多同步流形的概念，当节点自身具有多平衡点或多周期解的复杂动态行为时，基于分布式脉冲控制策略分别实现了网络的静态多同步和动态多同步；建立了量化控制策略实现了有限通信数据率下非线性多自主体网络的同步；提出了分层控制设计框架，为异构复杂多自主体网络协同提供了灵活的控制器设计方案，分别基于间歇通信机制和脉冲控制策略实现了编队包围和编队跟踪等多目标编队协同。项目执行期内发表学术论文共42篇，包括期刊论文26篇，会议论文16篇。其中SCI收录26篇。获湖北省自然科学奖一等奖1次，湖北省自然科学优秀学术论文二等奖1次。本项目拓展了复杂网络群集协调性的多样性内涵，对解决多功能型复杂多自主体网络的协调控制和多目标决策支持具有重要的理论意义和实用价值。