Distributed Optimization for Heterogenous Second⁃Order Multi⁃Agent Systems

A continuous⁃time distributed optimization was researched for second⁃order heterogeneous multi⁃agent systems.The aim of this study is to keep the velocities of all agents the same and make the velocities converge to the optimal value to minimize the sum of local cost functions.First,an effective distributed controller which only uses local information was designed.Then,the stability and optimization of the systems were verified.Finally,a simulation case was used to illustrate the analytical results.

基于Multi-Agent的无人机集群体系自主作战系统设计

针对无人集群自主作战体系设计中的关键问题,提出基于Multi-Agent的无人集群自主作战系统设计方法。建立无人集群各节点的Agent模型及其推演规则;对于仿真系统模块化和通用化的需求,设计系统互操作式接口和无人集群自主作战的交互关系;开展无人集群系统仿真推演验证。仿真结果表明,所提设计方案不仅能够有效开展并完成自主作战网络生成-集群演化-效能评估的全过程动态演示验证,而且能够通过重复随机试验进一步评估无人集群的协同作战效能,最后总结了集群协同作战的策略和经验。

Fixed-Time Cluster Optimization for Multi-Agent Systems Based on Piecewise Power-Law Design

Dear Editor,This letter focuses on the fixed-time(FXT)cluster optimization problem of first-order multi-agent systems(FOMASs)in an undirected network,in which the optimization objective is the sum of the objective functions of all clusters.A novel piecewise power-law control protocol with cooperative-competition relations is proposed.Furthermore,a sufficient condition is obtained to ensure that the FOMASs achieve the cluster consensus within an FXT.

Adaptive Consensus of Uncertain Multi-Agent Systems With Unified Prescribed Performance

Dear Editor,An adaptive consensus control algorithm for uncertain multi-agent systems(MAS),capable of guaranteeing unified prescribed performance,is presented in this letter.Unlike many existing prescribed performance related works,the developed control exhibits some features.Firstly,a distributed prescribed time observer is introduced so that not only each follower is able to estimate the leader’s signal within a predetermined time,but also the control design for each agent is independent with its neighbors.

Multi-Agent模式下的城市暴雨内涝应急决策方法研究

为厘清应对暴雨内涝灾害动态决策过程中决策主体、决策、决策方案等决策要素间的不确定关系,提出1种多主体(Multi-Agent)和贝叶斯决策网络(BDN)相结合的应急决策方法。首先分阶段构建“主体-任务”可视化网络,分析暴雨内涝灾害各应急阶段的主要任务和参与的决策主体;在考虑到决策要素间的动态不确定性可能造成决策风险的前提下,运用Multi-Agent和BDN方法探究各决策要素间的影响关系,以便进行方案集优选。研究结果表明:该方法具有实用性和现实意义,研究结果可为城市暴雨内涝灾害的应急决策提供理论参考。

基于免疫Agent的电力电缆线路故障检测系统

为区别故障信号、非故障信号,实现对电缆线路故障的准确检测,设计了基于免疫Agent的电力电缆线路故障检测系统。利用前端检测电路为故障分析模块、测距方式切换模块提供电量传输信号,完成系统前端检测装置的连接。按照免疫Agent检测原理提取电力电缆线路故障信号的特征,联合已获取信号对象,求解检测插值指标的具体数值。结合各级硬件设备结构,完成系统设计。实验结果表明,该系统可同时检测波频为10~20 Hz、40~50 Hz、60~70 Hz的故障信号与波频为20~30 Hz、30~40 Hz、50~60 Hz的非故障信号,可以在精准辨别故障与非故障信号的同时,实现对电力电缆线路故障的准确检测。

基于Agent人工智能的异构网络多重覆盖节点入侵检测系统设计

异构网络具有结构复杂、多重覆盖面积大等特征,使得网络入侵检测较为隐蔽,威胁网络运行的安全性;为此,对基于Agent人工智能的异构网络多重覆盖节点入侵检测系统进行了研究;通过检测Agent和通信Agent装设主机Agent,以Cisco Stealthwatch流量传感器作为异构网络传感器检测攻击行为,采用STM32L151RDT664位微控制器传输批量数据,由MAX3232芯片实现系统电平转化,实现硬件系统设计;软件部分设计入侵检测标准,采用传感器设备捕获网络实时数据,通过Agent技术解析异构网络协议并提取数据运行特征,综合考虑协议解析结果及与检测标准匹配度,实现异构网络多重覆盖节点入侵检测;经实验测试表明,基于Agent人工智能的异构网络多重覆盖节点入侵检测系统入侵行为的漏检率和入侵类型误检率的平均值仅为6%和5%,能够有效提高检测精度,减小检测误差。

基于多Agent传动关系的股市趋势预测

股市趋势预测是机器学习领域中一个具有挑战性的任务。由于一些因素对于股市的影响是动态且不确定的,导致股市趋势难以预测。针对已有方法在股市预测时存在的灵敏性差、适应力弱等问题,从快变量和慢变量的传动关系出发,利用Agent技术对股市中的快周期和慢周期进行联合建模,提出一种多Agent传动影响图(MATID)股市趋势预测方法。给出股市中快周期和慢周期的划分标准,并引入周期能量的概念;通过对相关趋势指标的特征融合,给出周期能量的量化计算方法;通过分析快周期和慢周期的动态作用过程,给出传动因子的表示方法;将快周期和慢周期分别对应成不同的Agent,利用多Agent影响图模型建模快周期和慢周期的传动过程;利用股市振子模型表示快Agent和慢Agent之间的传动效用,利用联合树的自动推理技术对股市趋势进行预测。在不同样本数量和不同股市趋势下进行实验,结果表明,与门控循环单元、S-LSTM和Hybrid-RNN预测方法相比,MATID方法预测精确率提升1.5%~7.0%,召回率提升5.4%~6.7%,F1值提升3.7%~6.2%,具有良好的灵敏性和适应力。

Early-Awareness Collision Avoidance in Optimal Multi-Agent Path Planning With Temporal Logic Specifications

Dear Editor,This letter investigates a multi-agent path planning problem in a road network with the requirement of avoiding collisions among all agents in the partitioned environment.We first abstract the agents to a set of transition systems,and construct a team transition system from these individual systems.A mechanism is designed for the team transition system to detect all collisions within the synthesized run.

竞争与合作视角下的多Agent强化学习研究进展

随着深度学习和强化学习研究取得长足的进展,多Agent强化学习已成为解决大规模复杂序贯决策问题的通用方法。为了推动该领域的发展,从竞争与合作的视角收集并总结近期相关的研究成果。该文介绍单Agent强化学习;分别介绍多Agent强化学习的基本理论框架——马尔可夫博弈以及扩展式博弈,并重点阐述了其在竞争、合作和混合三种场景下经典算法及其近期研究进展;讨论多Agent强化学习面临的核心挑战——环境的不稳定性,并通过一个例子对其解决思路进行总结与展望。

Task assignment in ground-to-air confrontation based on multiagent deep reinforcement learning

The scale of ground-to-air confrontation task assignments is large and needs to deal with many concurrent task assignments and random events.Aiming at the problems where existing task assignment methods are applied to ground-to-air confrontation,there is low efficiency in dealing with complex tasks,and there are interactive conflicts in multiagent systems.This study proposes a multiagent architecture based on a one-general agent with multiple narrow agents(OGMN)to reduce task assignment conflicts.Considering the slow speed of traditional dynamic task assignment algorithms,this paper proposes the proximal policy optimization for task assignment of general and narrow agents(PPOTAGNA)algorithm.The algorithm based on the idea of the optimal assignment strategy algorithm and combined with the training framework of deep reinforcement learning(DRL)adds a multihead attention mechanism and a stage reward mechanism to the bilateral band clipping PPO algorithm to solve the problem of low training efficiency.Finally,simulation experiments are carried out in the digital battlefield.The multiagent architecture based on OGMN combined with the PPO-TAGNA algorithm can obtain higher rewards faster and has a higher win ratio.By analyzing agent behavior,the efficiency,superiority and rationality of resource utilization of this method are verified.

Multi-Agent Deep Reinforcement Learning for Cross-Layer Scheduling in Mobile Ad-Hoc Networks

Due to the fading characteristics of wireless channels and the burstiness of data traffic,how to deal with congestion in Ad-hoc networks with effective algorithms is still open and challenging.In this paper,we focus on enabling congestion control to minimize network transmission delays through flexible power control.To effectively solve the congestion problem,we propose a distributed cross-layer scheduling algorithm,which is empowered by graph-based multi-agent deep reinforcement learning.The transmit power is adaptively adjusted in real-time by our algorithm based only on local information(i.e.,channel state information and queue length)and local communication(i.e.,information exchanged with neighbors).Moreover,the training complexity of the algorithm is low due to the regional cooperation based on the graph attention network.In the evaluation,we show that our algorithm can reduce the transmission delay of data flow under severe signal interference and drastically changing channel states,and demonstrate the adaptability and stability in different topologies.The method is general and can be extended to various types of topologies.

Multiple chemical warfare agent simulant decontamination by self-driven microplasma

Low-temperature plasma is a green and high-efficiency technology for chemical warfare agent(CWA)decontamination.However,traditional plasma devices suffer from the problems of highpower composition and large power-supply size,which limit their practical applications.In this paper,a self-driven microplasma decontamination system,induced by a dielectric-dielectric rotary triboelectric nanogenerator(dd-r TENG),was innovatively proposed for the decontamination of CWA simulants.The microplasma was characterized via electrical measurements,optical emission spectra and ozone concentration detection.With an output voltage of-3460 V,the dd-r TENG can successfully excite microplasma in air.Reactive species,such as OH,O(1D),Hαand O3were detected.With input average power of 0.116 W,the decontamination rate of 2-chloroethyl ethyl sulfide reached 100%within 3 min of plasma treatment,while the decontamination rates of malathion and dimethyl methylphosphonate reached(65.92±1.65)%and(60.88±1.92)%after 7 min of plasma treatment,respectively.In addition,the decontamination rates gradually decreased with the increase in the simulant concentrations.Typical products were identified and analyzed.This study demonstrates the broad spectrum and feasibility of the dd-r TENG-microplasma for CWA elimination,which provides significant guidance for their practical applications in the future.

Optimal Formation Control for Second-Order Multi-Agent Systems With Obstacle Avoidance

Dear Editor,The optimal formation control design problem is studied for a class of second-order multi-agent systems(MASs) with obstacle avoidance.Based on the actor-critic framework, an optimized formation controller is proposed by constructing a novel performance index function. Furthermore, the stability of MAS is proved by constructing the Lyapunov function. The simulation results are provided to depict the effectiveness of the proposed strategies.

Dynamic Target Enclosing Control Scheme for Multi-Agent Systems via a Signed Graph-Based Approach

Dear Editor,This letter investigates the target enclosing control problem of multi-agent systems.A signed graph-based control strategy is presented,where the agents are steered to enclose the dynamic target from both sides as they move.This is inspired by the phenomenon that signed networks exhibit bipartite clustering if the underlying graph is structurally balanced,so that the agents may naturally enclose the zero point from opposite sides(+and.)if proper controllers are applied.By adopting a distributed observer to estimate the information of dynamic target。

基于Multi-Agent技术的林产品供应链竞合研究

林产品贸易快速发展,随之而来的是竞争日益加剧,林产品供应链管理是林业产业发展的必然结果,供应链上企业间的竞合研究成为了热点。为此,搭建了基于Multi-Agent的林产品供应链,并在探讨核心制造商和销售商博弈收益的基础上,针对供应链中两者竞合关系进行了研究。通过对相应情况的建模仿真,在实验结果对比分析中发现制造商和销售商合作关系的影响因素有:初始合作比例、契约违约金和第三方监督,三者对合作行为有不同的影响,高水平的惩罚和监督可以有效抑制合作过程中的背叛行为。在此基础上,提出了培养信任文化、建立信誉评价平台和增设第三方监督主体的促进供应链合作的建议,为林产品供应链竞合研究提供参考。

基于多Agent的锂电池主动均衡策略控制仿真研究

【目的】针对锂电池的荷电状态均衡管理问题,提出一种基于多智能体的电池组荷电状态一致性均衡方案。【方法】首先,将多智能体控制策略引入电池管理的下垂控制中,实现了主动均衡电路拓扑下的自主均衡;其次,建立领航跟随者模型,利用参数已知的虚拟智能体使各个荷电状态不一致的电池的状态向其靠近,实现充放电模式下的荷电状态均衡;最后,对二阶多智能体荷电状态均衡控制策略进行仿真验证。【结果】实验结果表明,相比一阶均衡控制策略,自主均衡时间减少了43.02%,充电模式中均衡时间减少了16.13%,放电模式中均衡时间降低了32.90%。【结论】多智能体系统在电池的均衡管理中能够实现荷电状态的均衡,有效地降低了锂电池荷电状态到达一致性的收敛时间。

基于组合遗传算法的电力科技创新成果多agent集成仿真

电力科技成果普遍交叉,为了降低集成电力科技创新成果耗时与能量损耗,研究基于组合遗传算法的电力科技创新成果多agent集成仿真。将底层数据库中电力科技创新资源经多种接口调配至关联库、综合评价以及专家评价三大agent模块,通过结合遗传算法与蚁群算法优点的组合遗传算法,优化解决多agent集成任务分配问题,实现电力科技创新成果高效集成。实验表明,该算法的寻优能力高、求解速度快;可规范化、结构化地集成电力科技创新成果信息;具有耗时少、能量损耗低优势。为科技创新和成果转化提供新思路。

基于Agent的跨境电子商务课程智能教学系统

当前,跨境电子商务课程智能教学系统功能模块多为单向形式,可控教学覆盖范围较小,导致平均成绩优化比下降,为此提出对基于Agent的跨境电子商务课程智能教学系统的设计与分析。首先,构建通信接口,并接入驱动装置,完成系统硬件的设计;其次,采用多目标的方式,扩大可控教学覆盖范围;最后,设计多目标Agent跨境电子商务课程教学功能模块,关联跨境电子商务教学知识库,完成系统软件设计。测试结果表明:针对选定的6个班级进行测试对比,最终得出的平均成绩优化比均可以达到20%以上,说明在Agent的辅助下,文章设计的系统应用效果更佳,具有实际的应用价值。

A Robust Approach for Multi Classification-Based Intrusion Detection through Stacking Deep Learning Models

Intrusion detection is a predominant task that monitors and protects the network infrastructure.Therefore,many datasets have been published and investigated by researchers to analyze and understand the problem of intrusion prediction and detection.In particular,the Network Security Laboratory-Knowledge Discovery in Databases(NSL-KDD)is an extensively used benchmark dataset for evaluating intrusion detection systems(IDSs)as it incorporates various network traffic attacks.It is worth mentioning that a large number of studies have tackled the problem of intrusion detection using machine learning models,but the performance of these models often decreases when evaluated on new attacks.This has led to the utilization of deep learning techniques,which have showcased significant potential for processing large datasets and therefore improving detection accuracy.For that reason,this paper focuses on the role of stacking deep learning models,including convolution neural network(CNN)and deep neural network(DNN)for improving the intrusion detection rate of the NSL-KDD dataset.Each base model is trained on the NSL-KDD dataset to extract significant features.Once the base models have been trained,the stacking process proceeds to the second stage,where a simple meta-model has been trained on the predictions generated from the proposed base models.The combination of the predictions allows the meta-model to distinguish different classes of attacks and increase the detection rate.Our experimental evaluations using the NSL-KDD dataset have shown the efficacy of stacking deep learning models for intrusion detection.The performance of the ensemble of base models,combined with the meta-model,exceeds the performance of individual models.Our stacking model has attained an accuracy of 99%and an average F1-score of 93%for the multi-classification scenario.Besides,the training time of the proposed ensemble model is lower than the training time of benchmark techniques,demonstrating its efficiency and robustness.