多智能体网络中的动态事件驱动优化和一致性研究

The distributed optimization and control in multi-agent networks are hot research topics in the field of systems and control in recent years. In the practical application of multi-agent networks, there are still a series of problems such as limited communication bandwidth and insufficient onboard energy. In recent years, the event-triggered mechanism has become an effective way to solve the problems. Based on this background, this project will study the distributed optimization and consensus problem in multi-agent networks via the dynamic event-triggered mechanism. Firstly, we will design the dynamic event-triggered optimization algorithms to solve the distributed optimization problem in directed multi-agent networks. A comparison of the dynamic event-triggered optimization algorithm and the existing event-triggered optimization algorithm in information transmission pressure respect will be made. The effects on event-triggered optimization algorithm of the communication constraints will be discussed. Secondly, based on the existing dynamic event-triggered mechanism，we will investigate the consensus problem of the multi-agent systems with nonlinear dynamics. Combing the existing dynamic event-triggered mechanism with integral-type event-triggered mechanism, we will develop a novel event-triggered mechanism to derive the conditions for the dynamic event-triggered consensus problem of nonlinear multi-agent networks. The results will be applied to cope with formation control problem of the nonlinear multi-agent networks. This project will help to enrich and perfect the theoretical frameworks of distributed optimization and control, and provide an energy-saving strategy for the practical application of multi-agent networks.

多智能体网络中的优化和控制是当前系统和控制领域的研究热点。在多智能体网络的实际应用过程中，仍然存在通信带宽有限、机载能量不足等一系列问题。近年来，事件驱动机制策略已经成为解决这类问题的有效方式。本项目拟利用动态事件驱动机制来研究多智能体网络中的优化问题和一致性问题。首先，设计动态事件驱动优化算法，解决有向多智能体网络中的分布式优化问题，给出该算法与已有事件驱动优化算法在信息传输压力等方面的比较；探讨通信约束对事件驱动优化算法的影响。其次，基于已有动态事件驱动机制，研究具有非线性动态的多智能体网络的一致性问题；结合积分型事件驱动机制，设计新型动态事件驱动机制，推导非线性多智能体网络实现动态事件驱动一致性的条件；将所得结论应用于非线性多智能体网络的编队控制问题。本项目的研究不仅有助于丰富分布式优化和控制理论，同时也为多智能体网络的应用提供一种资源节省型策略。

项目的研究背景是，在多智能体网络的实际应用过程中，存在通信带宽有限、机载能量不足等一系列问题。近年来，事件驱动机制策略已经成为解决这类问题的有效方式。本项目的主要研究内容是利用事件驱动机制策略来解决多智能体网络中的优化问题和一致性问题，并进一步研究了多智能体系统的能控性问题。本项目的重要研究结果和关键数据，从三个方面展开。首先，设计了动态事件驱动优化算法，解决了有向多智能体网络中的分布式优化问题，给出了该算法与已有事件驱动优化算法在信息传输压力等方面的比较；揭示了通信约束对事件驱动优化算法的影响。其次，结合积分型事件驱动机制，推导了非线性多智能体网络实现事件驱动一致性的条件；利用动态事件驱动机制，解决了具有非线性动态的多智能体网络的一致性问题。最后，针对有限域上具有时变拓扑的多智能体系统，建立了这类系统的强目标可控性的图论判据，当系统的拓扑图满足一定条件时，证明了一个时变多智能体系统在任意有限域上都是强目标可控的，这部分结论对于事件驱动优化和一致性问题研究提供了重要的理论指导。本项目的研究意义，不仅丰富了分布式优化和控制理论，同时也为多智能体网络的应用提供了一种资源节省型策略。