变电站智能电子设备动态重构闭锁逻辑生成多代理系统

为保证变电站二次系统功能的完整和便于设备检修,智能电子设备(IED)动态重构技术逐渐得到关注。IED动态重构需要装载并解析IED配置描述(CID)文件,因管控主机保存的CID文件关联的是变电站设计时的系统结构,不能表征电力系统动态变化,为确保闭锁逻辑的正确性,需重新生成闭锁逻辑并更新相应的CID文件。利用多代理技术,管理动态重构管控主机、数据采集与监控(SCADA)主机、电力系统拓扑校验、变电站五防闭锁规则库生成软件等功能组件,构建多代理系统,完成电力系统拓扑校验和闭锁逻辑生成等任务的交互策略与协调流程,获得闭锁规则库更新CID文件,保证动态重构IED闭锁功能的正确性,能适应变电站的改造和升级,并提供更加灵活的闭锁规则管理。     

基于多代理系统的电动汽车协调充电策略

为最大化电力公司利益,设计了一种用于协调电动汽车充电的多代理系统,并在满足电动汽车车主充电需求及变压器容量限制的前提下,提出一种以负荷峰谷差最小为目标的分布式优化算法。利用分时电价算法初步优化后得到理想的充电时间区间,在充电区间内应用优化算法避免新的负荷尖峰,引入训练学习机制以使负荷曲线达到削峰填谷的效果。根据用户的驾驶习惯,采用蒙特卡洛方法模拟用户的充电需求,对电动汽车在无序充电、单次优化充电以及引入训练学习机制充电3种情况下的电网负荷进行了仿真分析。研究结果表明:单次优化可以避免负荷尖峰,但不能优化峰谷差;而引入训练学习机制后在减小峰谷差方面有显著作用,而且该分布式优化有更高的计算效率,适于实际应用。     

基于多层次LSTM网络的多智能体攻防效能动态预测模型

针对多智能体攻防体系存在多层次耦合、无规律涌现等特征而导致难以准确预测作战效能的问题,构建一种基于多层次长短时时间记忆(LSTM)网络的多智能体攻防效能动态预测模型。明确多智能体攻防的总体框架和作战流程,通过多主体NetLogo平台模拟红蓝智能体攻防对抗过程,以获取群体结构和作战效能在不同个体决策下的多层次演化数据。利用善于处理时序特征的LSTM网络来表征个体决策、群体结构和作战效能三层间的函数映射,并基于该映射关系进一步预测未来攻防作战效能与进程。上述建模方法已在多组仿真中证实了其可行性与有效性。实验结果表明,所建模型的作战效能预测误差仅在7%以内,对多智能体攻防的作战指挥和体系建设具有指导意义。     

基于滑模的二阶多智能体快速抗扰一致性

针对带有外部未知扰动的二阶多智能体系统的领导—跟随有限时间一致性问题，本文设计出一种带有时变增益的有限时间干扰观测器，用以实现对每个跟随智能体中未知扰动的快速估计，在此基础上，本文结合超螺旋积分滑模控制方法并利用邻居智能体的位置和速度信息设计一种快速抗扰一致性协议，该协议能够保证存在非线性动态的多智能体系统有限时间一致性控制并能抑制抖振现象.同时，利用李亚普诺夫函数进行了稳定性的证明.最后，通过Matlab数值仿真进一步验证了所提出协议的可行性.     

自主机器人多智能体软件架构及伴随行为机制

自主机器人是一类由计算机软件控制的信息物理系统，如何支持该类机器人在开放环境下的有效和协调运行是自主机器人控制软件（CSAR：Control Software of Autonomous Robot）研究与实践面临的一项重要挑战.本文基于组织理论的思想，采用Structure-in-5的组织架构模式，提出了基于多智能体的CSAR的软件架构MaRSA（Multi-agent Robotic Software Architecture），通过独立抽象CSAR的行为规划、分发、执行等软构件并显式加强这些构件间的交互，从而为自主机器人行为的有效规划和协调实施奠定架构基础；提出了基于MaRSA架构的伴随行为机制，从因果性、时序性和按需性三个方面建立了机器人观察行为和任务行为间的伴随关系，并基于分步规划和动态决策的思想设计并实现了伴随行为的自主决策算法DAAB（Decision Algorithm of Accompanying Behaviors）.论文分别在仿真环境和实际机器人环境下设计了对比性实验，结果表明，与主流的反应式行为决策算法和BDI式概率决策算法相比较，基于MaRSA和伴随行为机制的DAAB算法所生成的伴随行为规划在开放环境下具有可行性和更高效的执行效率.     

A multi-agent architecture for outsourcing SMEs manufacturing supply chain

In the present, global and competitive market customer's demands are very volatile. It is very difficult for small and medium size enterprises (SMEs) to satisfy their customer with quality product in reasonable price. Large firms are spending huge amount of their budget in generating technology development. To cope with this competitive market, usually, SMEs concentrate on their core expertise and outsource some non-core items or activities. The outsourcing decision (when, what, whom) is very crucial for small-scale industries. Realizing this scenario, in this article, an automated self-adaptive multi-agent system has been proposed. The proposed system will help SMEs to take an appropriate decision to mitigate the uncertainty in supply chain. The developed system exploits the “Outsourcing of operations” feature by its agents to conclude the manufacturing processes faster and reduce the idle time of certain machines with less human intervention. The communication protocol for agents has been described to give an insight of their communication. It was explained how various agents like outsourcing, material planning and supplier selection will come into action at different stages and take appropriate outsourcing decisions. Therefore, this multi-agent architecture will facilitate small scale manufacturing industries to execute their manufacturing processes and complex logistics issues efficiently. The execution of proposed architecture has been described in the simulated case study.     

Hierarchical correlated Q-learning for multi-layer optimal generation command dispatch

This paper presents a novel hierarchical correlated Q-learning (HCEQ) algorithm to solve the dynamic optimization of generation command dispatch (GCD) in the Automatic Generation Control (AGC). The GCD problem is to dynamically allocate the total AGC generation command from the central to each individual AGC generator. The proposed HCEQ is a novel multi-agent Q-learning algorithm based on the concept of correlated equilibrium point, and each AGC generator with an agent is to optimize its regulation participation factor and coordinate its decision with others for the overall GCD performance enhancement. In order to cope with the curse of dimensionality in the GCD problem with the increased number of AGC plants involved, a multi-layer optimum GCD framework is developed in this paper. In this hierarchical framework, the multiobjective design and a time-varying coordination factor have been formulated into the reward functions to improve the optimization efficiency and convergence of HCEQ. The application of the proposed approach has been fully verified on the China southern power grid (CSG) model to demonstrate its superior performance and dynamic optimization capability in various power system scenarios.     

Multi-formation control of nonlinear leader-following multi-agent systems

This paper deals with the multi-formation control problem for nonlinear leader-following multi-agent systems. Both the fixed topology case and the switching topology case are considered. The neighbor-based multi-formation control protocols are proposed under the assumption that for one subgroup, the total information received from other subgroups is zero. Then, based on the Lyapunov stability theory combined with the algebraic graph theory, sufficient conditions are established to ensure that the leader-following multi-agent systems with nonlinear dynamics can reach and maintain the desired multi-formation control. Finally, simulation examples are provided to illustrate the effectiveness of the theoretical results.     

Cooperative learning neural network output feedback control of uncertain nonlinear multi-agent systems under directed topologies

Without assuming that the communication topologies among the neural network (NN) weights are to be undirected and the states of each agent are measurable, the cooperative learning NN output feedback control is addressed for uncertain nonlinear multi-agent systems with identical structures in strict-feedback form. By establishing directed communication topologies among NN weights to share their learned knowledge, NNs with cooperative learning laws are employed to identify the uncertainties. By designing NN-based κ-filter observers to estimate the unmeasurable states, a new cooperative learning output feedback control scheme is proposed to guarantee that the system outputs can track nonidentical reference signals with bounded tracking errors. A simulation example is given to demonstrate the effectiveness of the theoretical results.     

Graph-balancing algorithms for average consensus over directed networks

Consensus strategies find extensive applications in coordination of robot groups and decision-making of agents. Since balanced graph plays an important role in the average consensus problem and many other coordination problems for directed communication networks, this work explores the conditions and algorithms for the digraph balancing problem. Based on the analysis of graph cycles, we prove that a digraph can be balanced if and only if the null space of its incidence matrix contains positive vectors. Then, based on this result and the corresponding analysis, two weight balance algorithms have been proposed, and the conditions for obtaining a unique balanced solution and a set of analytical results on weight balance problems have been introduced. Then, we point out the relationship between the weight balance problem and the features of the corresponding underlying Markov chain. Finally, two numerical examples are presented to verify the proposed algorithms.