检索结果-南通市图书馆

Intelligent Fast Cell Association Scheme Based on deep q-learning in Ultra-Dense Cellular Networks

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

China Communications 2021年第2期18卷 259-270页

作者： Jinhua Pan Lusheng Wang Hai Lin Zhiheng Zha Caihong Kai Key Laboratory of Knowledge Engineering with Big Data Ministry of Education.School of Computer Science and Information EngineeringHefei University of TechnologyHefei 230601China Anhui Province Key Laboratory of Industry Safety and Emergency Technology Hefei 230601China Key Laboratory of Aerospace Information Security and Trusted Computing Ministry of Education.School of Cyber Science and EngineeringWuhan UniversityWuhan 430072China

To support dramatically increased traffic loads,communication networks become *** cell association(CA)schemes are timeconsuming,forcing researchers to seek fast *** paper proposes a deep q-learning based scheme,whose main idea is to train a deep neural network(DNN)to calculate the q values of all the state-action pairs and the cell holding the maximum q value is *** the training stage,the intelligent agent continuously generates samples through the trial-anderror method to train the DNN until *** the application stage,state vectors of all the users are inputted to the trained DNN to quickly obtain a satisfied CA result of a scenario with the same BS locations and user *** demonstrate that the proposed scheme provides satisfied CA results in a computational time several orders of magnitudes shorter than traditional ***,performance metrics,such as capacity and fairness,can be guaranteed.

关键词： ultra-dense cellular networks(UDCN) cell association(CA) deep q-learning proportional fairness q-learning

Battlefield Situation Deduction and Maneuver Decision Using deep q-learning

在线全文

学校读者我要写书评

暂无评论

Battlefield Situation Deduction and Maneuver Decision Using ...

learning-based joint UAV trajectory and power allocation optimization for secure IoT networks

第40届中国控制会议

作者： Minghui Shi Xiwang Dong Liang Han qingdong Li Zhang Ren School of Automation Science and Electrical Engineering Science and Technology on Aircraft Control Laboratory Beihang University Key Laboratory of Dependable Service Computing in Cyber Physical Society (Chongqing University) Ministry of Education of ChinaChongqing University School of Sino-French Engineer Beihang University

As the pace and complexity of modern warfare accelerates, it is of great significance to apply intelligent technology to defense decision-making. Due to the high dynamics and randomness of the aircraft, traditional methods are difficult to solve the optimal control strategy. The characteristics of reinforcement learning match the difficulty of the problem. In this paper, the situation of hypersonic aircraft is deduced, and the deep reinforcement learning method is used to make autonomous penetration decisions in the reentry phase. The model of aircraft and environment is established, and the maneuvering decision-making model is established based on deep q-learning and its optimization algorithm. Through a large number of simulation training,this method can effectively give the real-time decision output of the agent and make a good prediction of the situation. It has the ability of short-range accurate operation and long-term planning and prediction. This method can improve the probability of successful penetration, and can be used as the decision-making basis of glider penetration.

关键词： Hypersonic aircraft situational deduction maneuver decision-making reinforcement learning deep q-learning

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

Digital Communications and Networks 2022年第4期8卷 415-421页

作者： Dan Deng Xingwang Li Varun Menon Md Jalil Piran Hui Chen Mian Ahmad Jan School of Information Engineering Guangzhou Panyu PolytechnicGuangzhou410630China School of Physics and Electronic Information Engineering Henan Polytechnic UniversityJiaozuo454150China Department of Computer Science and Engineering SCMS School of Engineering and TechnologyIndia Department of Computer Science and Engineering Sejong UniversitySouth Korea Department of Computer Science Abdul Wali Khan University MardanPakistan

Non-Orthogonal Multiplex Access(NOMA)can be deployed in Unmanned Aerial Vehicle(UAV)networks to improve spectrum *** to the broadcasting feature of NOMA-UAV networks,it is essential to focus on the security of the wireless *** paper focuses on maximizing the secrecy sum rate under the constraint of the achievable rate of the legitimate *** tackle the non-convexity optimization problem,a reinforcement learning-based alternative optimization algorithm is ***,with the help of successive convex approximations,the optimal power allocation scheme with a given UAV trajectory is obtained by using convex optimization ***,through plenty of explorations of the wireless environment,the q-learning networks approach the optimal location transition strategy of the UAV,even without the wireless channel state information.

关键词： Unmanned aerial vehicle(UAV) NOMA Reinforcement learning Secure communications deep q-learning

DqL-Based Intelligent Scheduling Algorithm for Automatic Driving in Massive MIMO V2I Scenarios

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

China Communications 2023年第3期20卷 18-26页

作者： Yong Liao Zisong Yin Zhijing Yang Xuanfan Shen School of Microelectronics and Communication Engineering Chongqing UniversityChongqing 400044China

Connected and autonomous vehicle(CAV)vehicle to infrastructure(V2I)scenarios have more stringent requirements on the communication rate,delay,and reliability of the Internet of vehicles(Io V).New radio vehicle to everything(NR-V2X)adopts link adaptation(LA)to improve the efficiency and reliability of road safety information *** order to solve the problem that the existing LA scheduling algorithms cannot adapt to the Doppler shift and complex fast time-varying channel in V2I scenario,resulting in low reliability of information transmission,this paper proposes a deep q-learning(DqL)-based massive multiple-input multiple-output(MIMO)LA scheduling algorithm for autonomous driving V2I *** algorithm combines deep neural network(DNN)with q-learning(qL)algorithm,which is used for joint scheduling of modulation and coding scheme(MCS)and space division multiplexing(SDM).The system simulation results show that the algorithm proposed in this paper can fully adapt to the different channel environment in the V2I scenario,and select the optimal MCS and SDM for the transmission of road safety information,thereby the accuracy of road safety information transmission is improved,collision accidents can be avoided,and bring a good autonomous driving experience.

关键词： NR autonomous driving V2I link adap-tation massive MIMO deep q-learning

deep Reinforcement learning-Based URLLC-Aware Task Offloading in Collaborative Vehicular Networks

在线全文

学校读者我要写书评

暂无评论

China Communications 2021年第7期18卷 134-146页

作者： Chao Pan Zhao Wang Zhenyu Zhou Xincheng Ren Hebei Key Laboratory of Power Internet of Things Technology North China Electric Power UniversityBeijing102206Baoding071003HebeiChina Shaanxi Key Laboratory of Intelligent Processing for Big Energy Data Yanan UniversityYanan 716000China

Collaborative vehicular networks is a key enabler to meet the stringent ultra-reliable and lowlatency communications(URLLC)requirements.A user vehicle(UV)dynamically optimizes task offloading by exploiting its collaborations with edge servers and vehicular fog servers(VFSs).However,the optimization of task offloading in highly dynamic collaborative vehicular networks faces several challenges such as URLLC guaranteeing,incomplete information,and dimensionality *** this paper,we first characterize URLLC in terms of queuing delay bound violation and high-order statistics of excess ***,a deep Reinforcement learning-based URLLCAware task offloading algorithM named DREAM is proposed to maximize the throughput of the UVs while satisfying the URLLC constraints in a besteffort *** with existing task offloading algorithms,DREAM achieves superior performance in throughput,queuing delay,and URLLC.

关键词： collaborative vehicular networks task of-floading URLLC awareness deep q-learning

在线全文

学校读者我要写书评

暂无评论

基于深度强化学习的热泵供热系统节能控制

建筑科学 2022年第12期38卷 1-6页

作者：秦浩森于震李太禄李立河北工业大学天津300131 中国建筑科学研究院有限公司北京100013

热泵供热系统广泛应用于住宅建筑,其优化控制对于提高需求侧的能源效率至关重要。基于模型的控制方法需要精确的建筑模型,而无模型控制方法前期效果较差,收敛速度较慢。针对这些问题,提出了1种基于deep q-learning及其改进算法的强化学... 详细信息

热泵供热系统广泛应用于住宅建筑,其优化控制对于提高需求侧的能源效率至关重要。基于模型的控制方法需要精确的建筑模型,而无模型控制方法前期效果较差,收敛速度较慢。针对这些问题,提出了1种基于deep q-learning及其改进算法的强化学习方法。该方法具有较快收敛速度,能够根据不同建筑环境自适应学习建模,在热舒适收益和能耗成本之间寻找平衡。北京市某近零能耗住宅建筑实际验证结果表明,该算法与基准策略相比综合收益提高15.3%。

关键词：强化学习 deep q-learning 控制策略暖通空调近零能耗建筑

Prediction of Boiler Combustion Energy Efficiency via deep Reinforcement learning

在线全文

学校读者我要写书评

暂无评论

Prediction of Boiler Combustion Energy Efficiency via Deep R...

一种MEC中的智能任务调度策略及其在ITS的应用研究

第40届中国控制会议

作者： Hui Jiang Ziyun Cai Tengfei Zhang Chen Peng College of Automation & College of Articial Intelligence Nanjing University of Posts and Telecommunications School of Mechatronic Engineering and Automation Shanghai University

Aiming at the problem of energy efficiency prediction during boiler combustion,a prediction method is *** boiler coal feed speed,steam value,feed water speed and power output are used as state variables,and the fuel utilization and rate of boiler combustion are used as the reward function to perform deep reinforcement learning *** experimental results show that the boiler combustion energy efficiency prediction system based on deep reinforcement learning proposed in this paper is effective and accurate,and performs well in the prediction process.

关键词： Boiler combustion deep q-learning Prediction

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

一种MEC中的智能任务调度策略及其在ITS的应用研究

作者：卢珊吉林大学

学位级别：硕士

随着智能移动终端设备的不断发展,新兴的人工智能应用(例如无人驾驶、虚拟现实、增强现实)的飞速普及,移动终端设备由于其有限的计算能力和电池容量而面临着巨大的挑战。传统的云计算具有强大的计算能力,但面对着远距离带来的响应时间... 详细信息

随着智能移动终端设备的不断发展,新兴的人工智能应用(例如无人驾驶、虚拟现实、增强现实)的飞速普及,移动终端设备由于其有限的计算能力和电池容量而面临着巨大的挑战。传统的云计算具有强大的计算能力,但面对着远距离带来的响应时间和传输带宽的限制,那些计算密集型与时延敏感型应用无法得到一个良好的反馈结果。为了应对这些问题,移动边缘计算(Mobile Edge Computing,MEC)随之出现在我们面前。MEC由于其具有强大的计算资源和存储资源,并且有着距离短、高带宽、低时延和低能耗等优点,已迅速成为了未来网络发展的主流方向。MEC中的主要难题就是解决任务调度问题,而任务调度的关键就是如何进行计算卸载,这也是本文研究工作的重点。目前关于用户进行计算卸载的许多研究没有考虑到用户对于服务器的选择问题。由于在实际情况下,服务器的有限计算资源和复杂的网络环境会导致诸多问题,其资源和负载不平衡等问题成为限制服务器可伸缩性和网络性能提高的关键因素。因此,本文考虑到了服务器资源的有限性,并在任务调度过程中产生的时延及能耗方面做了综合考量,设计了一种在MEC中的智能任务调度策略,并把该策略在智能交通系统中做了相关的应用研究。本文针对在MEC环境下的任务调度问题研究如下:本文首先阐述了MEC的研究现状,并分析了当前的研究重点。其次,介绍了本文使用的相关理论和技术。然后本文对MEC中的智能任务调度策略进行了模型搭建及设计,在多用户多服务器场景下,计算卸载是任务调度的一个重要研究方向。本文首先设计了一种基于优先级的任务调度策略,为移动用户的计算卸载提供了更多的选择方案,同时也使得计算资源的分配难度提升。服务器对于每一个申请卸载请求的用户都提供一个优先级等级,在用户面对多服务器可选择计算卸载的情况下,优先选择优先级等级最高的服务器进行卸载。本方案建立了在本地计算最大时延的约束下计算卸载总成功率最大化问题的优化模型。在提出的基于优先级的任务调度策略基础上,使用强化学习中的q-learning算法和深度学习中的卷积神经网络,进一步设计了一种基于deep q-learning的任务调度策略,对优先级算法的效率进行了提升。本文针对MEC中车载网络场景下的智能任务调度策略,模拟了一个街道场景并完成了实验。仿真实验结果表明,本文所设计的算法不仅能够显著提升计算任务的卸载成功率,还可以有效的降低任务调度过程中用户的时延与能耗。尤其是在边缘计算服务器资源紧张的情况下,本文所设计的智能任务调度策略能够充分利用服务器的计算资源与存储资源,提升整个系统的性能。

关键词：移动边缘计算任务调度计算卸载 deep q-learning

同方学位论文库评论

在线全文

同方学位论文库

学校读者我要写书评

暂无评论

基于强化学习的无人驾驶仿真研究

农业装备与车辆工程 2020年第6期58卷 102-106页

作者：孙嘉浩陈劲杰上海理工大学机械电子学院上海市200093

提出一种基于强化学习的无人驾驶仿真方案,采用deep q-learning算法,设置经验池来对驾驶策略进行学习,设计了控制策略和动作策略来控制虚拟环境下的驾驶仿真。在无人驾驶仿真平台TORCS上进行了仿真实验,对无人驾驶进行训练,训练结果验... 详细信息

提出一种基于强化学习的无人驾驶仿真方案,采用deep q-learning算法,设置经验池来对驾驶策略进行学习,设计了控制策略和动作策略来控制虚拟环境下的驾驶仿真。在无人驾驶仿真平台TORCS上进行了仿真实验,对无人驾驶进行训练,训练结果验证了该算法的有效性与可行性。该强化学习算法对无人驾驶仿真提供了可行方案的参考结论。

关键词：无人驾驶强化学习 deep q-learning 驾驶仿真 TORCS