Federated learning combines with fog computing to transform data sharing into model sharing,which solves the issues of data isolation and privacy disclosure in fog computing.However,existing studies focus on centraliz...Federated learning combines with fog computing to transform data sharing into model sharing,which solves the issues of data isolation and privacy disclosure in fog computing.However,existing studies focus on centralized single-layer aggregation federated learning architecture,which lack the consideration of cross-domain and asynchronous robustness of federated learning,and rarely integrate verification mechanisms from the perspective of incentives.To address the above challenges,we propose a Blockchain and Signcryption enabled Asynchronous Federated Learning(BSAFL)framework based on dual aggregation for cross-domain scenarios.In particular,we first design two types of signcryption schemes to secure the interaction and access control of collaborative learning between domains.Second,we construct a differential privacy approach that adaptively adjusts privacy budgets to ensure data privacy and local models'availability of intra-domain user.Furthermore,we propose an asynchronous aggregation solution that incorporates consensus verification and elastic participation using blockchain.Finally,security analysis demonstrates the security and privacy effectiveness of BSAFL,and the evaluation on real datasets further validates the high model accuracy and performance of BSAFL.展开更多
Considering the complexity of plant-wide optimization for large-scale industries, a distributed optimization framework to solve the profit optimization problem in ethylene whole process is proposed. To tackle the dela...Considering the complexity of plant-wide optimization for large-scale industries, a distributed optimization framework to solve the profit optimization problem in ethylene whole process is proposed. To tackle the delays arising from the residence time for materials passing through production units during the process with guaranteed constraint satisfaction, an asynchronous distributed parameter projection algorithm with gradient tracking method is introduced. Besides, the heavy ball momentum and Nesterov momentum are incorporated into the proposed algorithm in order to achieve double acceleration properties. The experimental results show that the proposed asynchronous algorithm can achieve a faster convergence compared with the synchronous algorithm.展开更多
To address the shortcomings of traditional Genetic Algorithm (GA) in multi-agent path planning, such as prolonged planning time, slow convergence, and solution instability, this paper proposes an Asynchronous Genetic ...To address the shortcomings of traditional Genetic Algorithm (GA) in multi-agent path planning, such as prolonged planning time, slow convergence, and solution instability, this paper proposes an Asynchronous Genetic Algorithm (AGA) to solve multi-agent path planning problems effectively. To enhance the real-time performance and computational efficiency of Multi-Agent Systems (MAS) in path planning, the AGA incorporates an Equal-Size Clustering Algorithm (ESCA) based on the K-means clustering method. The ESCA divides the primary task evenly into a series of subtasks, thereby reducing the gene length in the subsequent GA process. The algorithm then employs GA to solve each subtask sequentially. To evaluate the effectiveness of the proposed method, a simulation program was designed to perform path planning for 100 trajectories, and the results were compared with those of State-Of-The-Art (SOTA) methods. The simulation results demonstrate that, although the solutions provided by AGA are suboptimal, it exhibits significant advantages in terms of execution speed and solution stability compared to other algorithms.展开更多
Photomechanics is a crucial branch of solid mechanics.The localization of point targets constitutes a fundamental problem in optical experimental mechanics,with extensive applications in various missions of unmanned a...Photomechanics is a crucial branch of solid mechanics.The localization of point targets constitutes a fundamental problem in optical experimental mechanics,with extensive applications in various missions of unmanned aerial vehicles.Localizing moving targets is crucial for analyzing their motion characteristics and dynamic properties.Reconstructing the trajectories of points from asynchronous cameras is a significant challenge.It encompasses two coupled sub-problems:Trajectory reconstruction and camera synchronization.Present methods typically address only one of these sub-problems individually.This paper proposes a 3D trajectory reconstruction method for point targets based on asynchronous cameras,simultaneously solving both sub-problems.Firstly,we extend the trajectory intersection method to asynchronous cameras to resolve the limitation of traditional triangulation that requires camera synchronization.Secondly,we develop models for camera temporal information and target motion,based on imaging mechanisms and target dynamics characteristics.The parameters are optimized simultaneously to achieve trajectory reconstruction without accurate time parameters.Thirdly,we optimize the camera rotations alongside the camera time information and target motion parameters,using tighter and more continuous constraints on moving points.The reconstruction accuracy is significantly improved,especially when the camera rotations are inaccurate.Finally,the simulated and real-world experimental results demonstrate the feasibility and accuracy of the proposed method.The real-world results indicate that the proposed algorithm achieved a localization error of 112.95 m at an observation distance range of 15-20 km.展开更多
Federated learning combined with edge computing has greatly facilitated transportation in real-time applications such as intelligent traffic sys-tems.However,synchronous federated learning is in-efficient in terms of ...Federated learning combined with edge computing has greatly facilitated transportation in real-time applications such as intelligent traffic sys-tems.However,synchronous federated learning is in-efficient in terms of time and convergence speed,mak-ing it unsuitable for high real-time requirements.To address these issues,this paper proposes an Adap-tive Waiting time Asynchronous Federated Learn-ing(AWTAFL)based on Dueling Double Deep Q-Network(D3QN).The server dynamically adjusts the waiting time using the D3QN algorithm based on the current task progress and energy consumption,aim-ing to accelerate convergence and save energy.Addi-tionally,this paper presents a new federated learning global aggregation scheme,where the central server performs weighted aggregation based on the freshness and contribution of client parameters.Experimen-tal simulations demonstrate that the proposed algo-rithm significantly reduces the convergence time while ensuring model quality and effectively reducing en-ergy consumption in asynchronous federated learning.Furthermore,the improved global aggregation update method enhances training stability and reduces oscil-lations in the global model convergence.展开更多
Dense-array ambient noise tomography is a powerful tool for achieving high-resolution subsurface imag-ing,significantly impacting geohazard prevention and control.Conventional dense-array studies,how-ever,require simu...Dense-array ambient noise tomography is a powerful tool for achieving high-resolution subsurface imag-ing,significantly impacting geohazard prevention and control.Conventional dense-array studies,how-ever,require simultaneous observations of numerous stations for extensive coverage.To conduct a comprehensive karst feature investigation with limited stations,we designed a new synchronous-asyn-chronous observation system that facilitates dense array observations.We conducted two rounds of asynchronous observations,each lasting approximately 24 h,in combination with synchronous backbone stations.We achieved wide-ranging coverage of the study area utilizing 197 nodal receivers,with an average station spacing of 7 m.The beamforming results revealed distinct variations in the noise source distributions between day and night.We estimated the source strength in the stationary phase zone and used a weighting scheme for stacking the cross-correlation functions(C ^(1) functions)to suppress the influ-ence of nonuniform noise source distributions.The weights were derived from the similarity coefficients between multicomponent C^(1)functions related to Rayleigh waves.We employed the cross-correlation of C ^(1) functions(C^(2)methods)to obtain the empirical Green’s functions between asynchronous stations.To eliminate artifacts in C ^(2) functions from higher-mode surface waves in C^(1)functions,we filtered the C^(1)functions on the basis of different particle motions linked to multimode Rayleigh waves.The dispersion measurements of Rayleigh waves obtained from both the C^(1)and C^(2)functions were utilized in surface wave tomography.The inverted three-dimensional(3D)shear-wave(S-wave)velocity model reveals two significant low-velocity zones at depths ranging from 40 to 60 m,which align well with the karst caves found in the drilling data.The method of short-term synchronous-asynchronous ambient noise tomography shows promise as a cost-effective and efficient approach for urban geohazard investigations.展开更多
This research focuses on detecting faults in flight vehicles with unstable subsystems operating asynchronously.By accounting for asynchronous switching,a switched model is established,and filters for fault detection(F...This research focuses on detecting faults in flight vehicles with unstable subsystems operating asynchronously.By accounting for asynchronous switching,a switched model is established,and filters for fault detection(FD)in unstable subsystems are developed.The FD challenge is then transformed into an H∞filtering issue.Utilizing the multiple discontinuous Lyapunov function(MDLF)approach and the mode-dependent average dwell time(MDADT)method,sufficient conditions are derived to ensure stability during both fast and slow switching.Furthermore,the existence and solutions for FD filters are provided through linear matrix inequalities(LMIs).The simulation outcomes demonstrated the excellent performance of the developed method in studied cases.展开更多
In vehicle edge computing(VEC),asynchronous federated learning(AFL)is used,where the edge receives a local model and updates the global model,effectively reducing the global aggregation latency.Due to different amount...In vehicle edge computing(VEC),asynchronous federated learning(AFL)is used,where the edge receives a local model and updates the global model,effectively reducing the global aggregation latency.Due to different amounts of local data,computing capabilities and locations of the vehicles,renewing the global model with same weight is inappropriate.The above factors will affect the local calculation time and upload time of the local model,and the vehicle may also be affected by Byzantine attacks,leading to the deterioration of the vehicle data.However,based on deep reinforcement learning(DRL),we can consider these factors comprehensively to eliminate vehicles with poor performance as much as possible and exclude vehicles that have suffered Byzantine attacks before AFL.At the same time,when aggregating AFL,we can focus on those vehicles with better performance to improve the accuracy and safety of the system.In this paper,we proposed a vehicle selection scheme based on DRL in VEC.In this scheme,vehicle’s mobility,channel conditions with temporal variations,computational resources with temporal variations,different data amount,transmission channel status of vehicles as well as Byzantine attacks were taken into account.Simulation results show that the proposed scheme effectively improves the safety and accuracy of the global model.展开更多
The clay mineral flocculation encapsulation poses a major technical challenge in the field of fine mineral separation.Enhancing the ability to separate clay minerals from target mineral surfaces is key to addressing t...The clay mineral flocculation encapsulation poses a major technical challenge in the field of fine mineral separation.Enhancing the ability to separate clay minerals from target mineral surfaces is key to addressing this issue.In the flotation process of ultrafine hematite,sodium polyacrylate(PAAS)was used as a selective flocculant for hematite,polyaluminum chloride(PAC)as a flocculant for kaolinite and chlorite,and sodium oleate(NaOL)as the collector to achieve asynchronous flocculation flotation.This study examines the flotation separation performance and validates it through experiments on actual mineral samples.The results indicate that with PAAS and PAC dosages of 1.25 and 50 mg·L^(-1),respectively,the iron grade and recovery of the actual mineral samples increased by 9.39%and 7.97%.Through Zeta potential,XPS analysis,infrared spectroscopy,and total organic carbon(TOC)testing,the study reveals the microscopic interaction mechanisms of different flocculants with minerals,providing insights for the clean and efficient utilization of ultrafine mineral resources.展开更多
To provide diversified services in the intelligent transportation systems,smart vehicles will generate unprecedented amounts of data every day.Due to data security and user privacy issues,Federated Learning(FL)is cons...To provide diversified services in the intelligent transportation systems,smart vehicles will generate unprecedented amounts of data every day.Due to data security and user privacy issues,Federated Learning(FL)is considered a potential solution to ensure privacy-preserving in data sharing.However,there are still many challenges to applying the traditional synchronous FL directly in the Internet of Vehicles(Io V),such as unreliable communications and malicious attacks.In this paper,we propose a Directed Acyclic Graph(DAG)based Swarm Learning(DSL),which integrates edge computing,FL,and blockchain technologies to provide secure data sharing and model training in Io Vs.To deal with the high mobility of vehicles,the dynamic vehicle association algorithm is introduced,which could optimize the connections between vehicles and road side units to improve the training efficiency.Moreover,to enhance the anti-attack property of the DSL algorithm,a malicious attack detection method is adopted,which could recognize malicious vehicles by the site confirmation rate.Furthermore,an accuracy-based reward mechanism is developed to promote vehicles to participate in the model training with honest behaviors.Finally,simulation results demonstrate that the proposed DSL algorithm could achieve better performance in terms of model accuracy,convergence rates and security compared with existing algorithms.展开更多
In Software-Defined Networks(SDNs),determining how to efficiently achieve Quality of Service(QoS)-aware routing is challenging but critical for significantly improving the performance of a network,where the metrics of...In Software-Defined Networks(SDNs),determining how to efficiently achieve Quality of Service(QoS)-aware routing is challenging but critical for significantly improving the performance of a network,where the metrics of QoS can be defined as,for example,average latency,packet loss ratio,and throughput.The SDN controller can use network statistics and a Deep Reinforcement Learning(DRL)method to resolve this challenge.In this paper,we formulate dynamic routing in an SDN as a Markov decision process and propose a DRL algorithm called the Asynchronous Advantage Actor-Critic QoS-aware Routing Optimization Mechanism(AQROM)to determine routing strategies that balance the traffic loads in the network.AQROM can improve the QoS of the network and reduce the training time via dynamic routing strategy updates;that is,the reward function can be dynamically and promptly altered based on the optimization objective regardless of the network topology and traffic pattern.AQROM can be considered as one-step optimization and a black-box routing mechanism in high-dimensional input and output sets for both discrete and continuous states,and actions with respect to the operations in the SDN.Extensive simulations were conducted using OMNeT++and the results demonstrated that AQROM 1)achieved much faster and stable convergence than the Deep Deterministic Policy Gradient(DDPG)and Advantage Actor-Critic(A2C),2)incurred a lower packet loss ratio and latency than Open Shortest Path First(OSPF),DDPG,and A2C,and 3)resulted in higher and more stable throughput than OSPF,DDPG,and A2C.展开更多
基金supported in part by the National Key Research and Development Program of China under Grant No.2021YFB3101100in part by the National Natural Science Foundation of China under Grant 62272123,62272102,62272124+2 种基金in part by the Project of High-level Innovative Talents of Guizhou Province under Grant[2020]6008in part by the Science and Technology Program of Guizhou Province under Grant No.[2020]5017,No.[2022]065in part by the Guangxi Key Laboratory of Cryptography and Information Security under Grant GCIS202105。
文摘Federated learning combines with fog computing to transform data sharing into model sharing,which solves the issues of data isolation and privacy disclosure in fog computing.However,existing studies focus on centralized single-layer aggregation federated learning architecture,which lack the consideration of cross-domain and asynchronous robustness of federated learning,and rarely integrate verification mechanisms from the perspective of incentives.To address the above challenges,we propose a Blockchain and Signcryption enabled Asynchronous Federated Learning(BSAFL)framework based on dual aggregation for cross-domain scenarios.In particular,we first design two types of signcryption schemes to secure the interaction and access control of collaborative learning between domains.Second,we construct a differential privacy approach that adaptively adjusts privacy budgets to ensure data privacy and local models'availability of intra-domain user.Furthermore,we propose an asynchronous aggregation solution that incorporates consensus verification and elastic participation using blockchain.Finally,security analysis demonstrates the security and privacy effectiveness of BSAFL,and the evaluation on real datasets further validates the high model accuracy and performance of BSAFL.
基金supported by National Key Research and Development Program of China(2022YFB3305900)National Natural Science Foundation of China(62394343,62394345)+1 种基金Major Science and Technology Projects of Longmen Laboratory(NO.LMZDXM202206)Shanghai Rising-Star Program under Grant 24QA2706100.
文摘Considering the complexity of plant-wide optimization for large-scale industries, a distributed optimization framework to solve the profit optimization problem in ethylene whole process is proposed. To tackle the delays arising from the residence time for materials passing through production units during the process with guaranteed constraint satisfaction, an asynchronous distributed parameter projection algorithm with gradient tracking method is introduced. Besides, the heavy ball momentum and Nesterov momentum are incorporated into the proposed algorithm in order to achieve double acceleration properties. The experimental results show that the proposed asynchronous algorithm can achieve a faster convergence compared with the synchronous algorithm.
文摘To address the shortcomings of traditional Genetic Algorithm (GA) in multi-agent path planning, such as prolonged planning time, slow convergence, and solution instability, this paper proposes an Asynchronous Genetic Algorithm (AGA) to solve multi-agent path planning problems effectively. To enhance the real-time performance and computational efficiency of Multi-Agent Systems (MAS) in path planning, the AGA incorporates an Equal-Size Clustering Algorithm (ESCA) based on the K-means clustering method. The ESCA divides the primary task evenly into a series of subtasks, thereby reducing the gene length in the subsequent GA process. The algorithm then employs GA to solve each subtask sequentially. To evaluate the effectiveness of the proposed method, a simulation program was designed to perform path planning for 100 trajectories, and the results were compared with those of State-Of-The-Art (SOTA) methods. The simulation results demonstrate that, although the solutions provided by AGA are suboptimal, it exhibits significant advantages in terms of execution speed and solution stability compared to other algorithms.
基金supported by the Hunan Provin〓〓cial Natural Science Foundation for Excellent Young Scholars(Grant No.2023JJ20045)the National Natural Science Foundation of China(Grant No.12372189)。
文摘Photomechanics is a crucial branch of solid mechanics.The localization of point targets constitutes a fundamental problem in optical experimental mechanics,with extensive applications in various missions of unmanned aerial vehicles.Localizing moving targets is crucial for analyzing their motion characteristics and dynamic properties.Reconstructing the trajectories of points from asynchronous cameras is a significant challenge.It encompasses two coupled sub-problems:Trajectory reconstruction and camera synchronization.Present methods typically address only one of these sub-problems individually.This paper proposes a 3D trajectory reconstruction method for point targets based on asynchronous cameras,simultaneously solving both sub-problems.Firstly,we extend the trajectory intersection method to asynchronous cameras to resolve the limitation of traditional triangulation that requires camera synchronization.Secondly,we develop models for camera temporal information and target motion,based on imaging mechanisms and target dynamics characteristics.The parameters are optimized simultaneously to achieve trajectory reconstruction without accurate time parameters.Thirdly,we optimize the camera rotations alongside the camera time information and target motion parameters,using tighter and more continuous constraints on moving points.The reconstruction accuracy is significantly improved,especially when the camera rotations are inaccurate.Finally,the simulated and real-world experimental results demonstrate the feasibility and accuracy of the proposed method.The real-world results indicate that the proposed algorithm achieved a localization error of 112.95 m at an observation distance range of 15-20 km.
基金supported by the National Natural Science Foundation of China(62371082)Guangxi Science and Technology Project(AB24010317)+1 种基金Science and Technology Project of Chongqing Education Commission(KJZD-K202400606)Natural Science Foundation of Chongqing(CSTB2023NSCQ-MSX0726,CSTB2023NSCQ-LZX0014).
文摘Federated learning combined with edge computing has greatly facilitated transportation in real-time applications such as intelligent traffic sys-tems.However,synchronous federated learning is in-efficient in terms of time and convergence speed,mak-ing it unsuitable for high real-time requirements.To address these issues,this paper proposes an Adap-tive Waiting time Asynchronous Federated Learn-ing(AWTAFL)based on Dueling Double Deep Q-Network(D3QN).The server dynamically adjusts the waiting time using the D3QN algorithm based on the current task progress and energy consumption,aim-ing to accelerate convergence and save energy.Addi-tionally,this paper presents a new federated learning global aggregation scheme,where the central server performs weighted aggregation based on the freshness and contribution of client parameters.Experimen-tal simulations demonstrate that the proposed algo-rithm significantly reduces the convergence time while ensuring model quality and effectively reducing en-ergy consumption in asynchronous federated learning.Furthermore,the improved global aggregation update method enhances training stability and reduces oscil-lations in the global model convergence.
基金supported by the National Natural Science Foundation of China(41830103)the Project of Nanjing Center of China Geological Survey(DD20190281).
文摘Dense-array ambient noise tomography is a powerful tool for achieving high-resolution subsurface imag-ing,significantly impacting geohazard prevention and control.Conventional dense-array studies,how-ever,require simultaneous observations of numerous stations for extensive coverage.To conduct a comprehensive karst feature investigation with limited stations,we designed a new synchronous-asyn-chronous observation system that facilitates dense array observations.We conducted two rounds of asynchronous observations,each lasting approximately 24 h,in combination with synchronous backbone stations.We achieved wide-ranging coverage of the study area utilizing 197 nodal receivers,with an average station spacing of 7 m.The beamforming results revealed distinct variations in the noise source distributions between day and night.We estimated the source strength in the stationary phase zone and used a weighting scheme for stacking the cross-correlation functions(C ^(1) functions)to suppress the influ-ence of nonuniform noise source distributions.The weights were derived from the similarity coefficients between multicomponent C^(1)functions related to Rayleigh waves.We employed the cross-correlation of C ^(1) functions(C^(2)methods)to obtain the empirical Green’s functions between asynchronous stations.To eliminate artifacts in C ^(2) functions from higher-mode surface waves in C^(1)functions,we filtered the C^(1)functions on the basis of different particle motions linked to multimode Rayleigh waves.The dispersion measurements of Rayleigh waves obtained from both the C^(1)and C^(2)functions were utilized in surface wave tomography.The inverted three-dimensional(3D)shear-wave(S-wave)velocity model reveals two significant low-velocity zones at depths ranging from 40 to 60 m,which align well with the karst caves found in the drilling data.The method of short-term synchronous-asynchronous ambient noise tomography shows promise as a cost-effective and efficient approach for urban geohazard investigations.
基金the National Natural Science Foundation of China(Grant Nos.62303380,62176214,62101590,62003268)the Aeronautical Science Foundation of China(Grant No.201907053001).
文摘This research focuses on detecting faults in flight vehicles with unstable subsystems operating asynchronously.By accounting for asynchronous switching,a switched model is established,and filters for fault detection(FD)in unstable subsystems are developed.The FD challenge is then transformed into an H∞filtering issue.Utilizing the multiple discontinuous Lyapunov function(MDLF)approach and the mode-dependent average dwell time(MDADT)method,sufficient conditions are derived to ensure stability during both fast and slow switching.Furthermore,the existence and solutions for FD filters are provided through linear matrix inequalities(LMIs).The simulation outcomes demonstrated the excellent performance of the developed method in studied cases.
基金supported in part by the National Natural Science Foundation of China(No.61701197)in part by the National Key Research and Development Program of China(No.2021YFA1000500(4))in part by the 111 Project(No.B23008).
文摘In vehicle edge computing(VEC),asynchronous federated learning(AFL)is used,where the edge receives a local model and updates the global model,effectively reducing the global aggregation latency.Due to different amounts of local data,computing capabilities and locations of the vehicles,renewing the global model with same weight is inappropriate.The above factors will affect the local calculation time and upload time of the local model,and the vehicle may also be affected by Byzantine attacks,leading to the deterioration of the vehicle data.However,based on deep reinforcement learning(DRL),we can consider these factors comprehensively to eliminate vehicles with poor performance as much as possible and exclude vehicles that have suffered Byzantine attacks before AFL.At the same time,when aggregating AFL,we can focus on those vehicles with better performance to improve the accuracy and safety of the system.In this paper,we proposed a vehicle selection scheme based on DRL in VEC.In this scheme,vehicle’s mobility,channel conditions with temporal variations,computational resources with temporal variations,different data amount,transmission channel status of vehicles as well as Byzantine attacks were taken into account.Simulation results show that the proposed scheme effectively improves the safety and accuracy of the global model.
基金funded by the National Natural Science Foundation of China(No.52374265)the Central Guided Local Science and Technology Development Funding Program(No.236Z4106G)+1 种基金the Natural Science Foundation of Hebei Province(No.E2022209108)Key Projects of Hebei Provincial Department of Education(No.ZD2022059)。
文摘The clay mineral flocculation encapsulation poses a major technical challenge in the field of fine mineral separation.Enhancing the ability to separate clay minerals from target mineral surfaces is key to addressing this issue.In the flotation process of ultrafine hematite,sodium polyacrylate(PAAS)was used as a selective flocculant for hematite,polyaluminum chloride(PAC)as a flocculant for kaolinite and chlorite,and sodium oleate(NaOL)as the collector to achieve asynchronous flocculation flotation.This study examines the flotation separation performance and validates it through experiments on actual mineral samples.The results indicate that with PAAS and PAC dosages of 1.25 and 50 mg·L^(-1),respectively,the iron grade and recovery of the actual mineral samples increased by 9.39%and 7.97%.Through Zeta potential,XPS analysis,infrared spectroscopy,and total organic carbon(TOC)testing,the study reveals the microscopic interaction mechanisms of different flocculants with minerals,providing insights for the clean and efficient utilization of ultrafine mineral resources.
基金supported in part by the National Natural Science Foundation of China(NSFC)under Grant 62371082,61831002,and 62001076in part by the Natural Science Foundation of Chongqing under Grant CSTB2023NSCQ-MSX0726 and cstc2020jcyj-msxmX0878。
文摘To provide diversified services in the intelligent transportation systems,smart vehicles will generate unprecedented amounts of data every day.Due to data security and user privacy issues,Federated Learning(FL)is considered a potential solution to ensure privacy-preserving in data sharing.However,there are still many challenges to applying the traditional synchronous FL directly in the Internet of Vehicles(Io V),such as unreliable communications and malicious attacks.In this paper,we propose a Directed Acyclic Graph(DAG)based Swarm Learning(DSL),which integrates edge computing,FL,and blockchain technologies to provide secure data sharing and model training in Io Vs.To deal with the high mobility of vehicles,the dynamic vehicle association algorithm is introduced,which could optimize the connections between vehicles and road side units to improve the training efficiency.Moreover,to enhance the anti-attack property of the DSL algorithm,a malicious attack detection method is adopted,which could recognize malicious vehicles by the site confirmation rate.Furthermore,an accuracy-based reward mechanism is developed to promote vehicles to participate in the model training with honest behaviors.Finally,simulation results demonstrate that the proposed DSL algorithm could achieve better performance in terms of model accuracy,convergence rates and security compared with existing algorithms.
基金fully supported by GUET Excellent Graduate Thesis Program(Grant No.19YJPYBS03)Innovation Project of Guangxi Graduate Education(Grant No.YCBZ2022109)New Technology Research University Cooperation Project of the 34th Research Institute of China Electronics Technology Group Corporation,2021(Grant No.SF2126007)。
文摘In Software-Defined Networks(SDNs),determining how to efficiently achieve Quality of Service(QoS)-aware routing is challenging but critical for significantly improving the performance of a network,where the metrics of QoS can be defined as,for example,average latency,packet loss ratio,and throughput.The SDN controller can use network statistics and a Deep Reinforcement Learning(DRL)method to resolve this challenge.In this paper,we formulate dynamic routing in an SDN as a Markov decision process and propose a DRL algorithm called the Asynchronous Advantage Actor-Critic QoS-aware Routing Optimization Mechanism(AQROM)to determine routing strategies that balance the traffic loads in the network.AQROM can improve the QoS of the network and reduce the training time via dynamic routing strategy updates;that is,the reward function can be dynamically and promptly altered based on the optimization objective regardless of the network topology and traffic pattern.AQROM can be considered as one-step optimization and a black-box routing mechanism in high-dimensional input and output sets for both discrete and continuous states,and actions with respect to the operations in the SDN.Extensive simulations were conducted using OMNeT++and the results demonstrated that AQROM 1)achieved much faster and stable convergence than the Deep Deterministic Policy Gradient(DDPG)and Advantage Actor-Critic(A2C),2)incurred a lower packet loss ratio and latency than Open Shortest Path First(OSPF),DDPG,and A2C,and 3)resulted in higher and more stable throughput than OSPF,DDPG,and A2C.
