The smart grid utilizes the demand side management technology to motivate energy users towards cutting demand during peak power consumption periods, which greatly improves the operation efficiency of the power grid. H...The smart grid utilizes the demand side management technology to motivate energy users towards cutting demand during peak power consumption periods, which greatly improves the operation efficiency of the power grid. However, as the number of energy users participating in the smart grid continues to increase, the demand side management strategy of individual agent is greatly affected by the dynamic strategies of other agents. In addition, the existing demand side management methods, which need to obtain users’ power consumption information,seriously threaten the users’ privacy. To address the dynamic issue in the multi-microgrid demand side management model, a novel multi-agent reinforcement learning method based on centralized training and decentralized execution paradigm is presented to mitigate the damage of training performance caused by the instability of training experience. In order to protect users’ privacy, we design a neural network with fixed parameters as the encryptor to transform the users’ energy consumption information from low-dimensional to high-dimensional and theoretically prove that the proposed encryptor-based privacy preserving method will not affect the convergence property of the reinforcement learning algorithm. We verify the effectiveness of the proposed demand side management scheme with the real-world energy consumption data of Xi’an, Shaanxi, China. Simulation results show that the proposed method can effectively improve users’ satisfaction while reducing the bill payment compared with traditional reinforcement learning(RL) methods(i.e., deep Q learning(DQN), deep deterministic policy gradient(DDPG),QMIX and multi-agent deep deterministic policy gradient(MADDPG)). The results also demonstrate that the proposed privacy protection scheme can effectively protect users’ privacy while ensuring the performance of the algorithm.展开更多
Controllers play a critical role in software-defined networking(SDN).However,existing singlecontroller SDN architectures are vulnerable to single-point failures,where a controller's capacity can be saturated by fl...Controllers play a critical role in software-defined networking(SDN).However,existing singlecontroller SDN architectures are vulnerable to single-point failures,where a controller's capacity can be saturated by flooded flow requests.In addition,due to the complicated interactions between applications and controllers,the flow setup latency is relatively large.To address the above security and performance issues of current SDN controllers,we propose distributed rule store(DRS),a new multi-controller architecture for SDNs.In DRS,the controller caches the flow rules calculated by applications,and distributes these rules to multiple controller instances.Each controller instance holds only a subset of all rules,and periodically checks the consistency of flow rules with each other.Requests from switches are distributed among multiple controllers,in order to mitigate controller capacity saturation attack.At the same time,when rules at one controller are maliciously modified,they can be detected and recovered in time.We implement DRS based on Floodlight and evaluate it with extensive emulation.The results show that DRS can effectively maintain a consistently distributed rule store,and at the same time can achieve a shorter flow setup time and a higher processing throughput,compared with ONOS and Floodlight.展开更多
Every human being looks different in one way or the other.That’s the work of genetic variations,the ultimate driving force for evolution as well as the cause for many human diseases.Mapping human genetic variants rev...Every human being looks different in one way or the other.That’s the work of genetic variations,the ultimate driving force for evolution as well as the cause for many human diseases.Mapping human genetic variants reveals global genetic diversity,and pinpoints causal variants behind genetic disorders.展开更多
Graves' disease,the production of thyroid-stimulating hormone receptor-stimulating antibodies leading to hyperthyroidism,is one of the most common forms of human autoimmune disease.It is widely agreed that complex...Graves' disease,the production of thyroid-stimulating hormone receptor-stimulating antibodies leading to hyperthyroidism,is one of the most common forms of human autoimmune disease.It is widely agreed that complex diseases are not controlled simply by an individual gene or DNA variation but by their combination.Single nucleotide polymorphisms(SNPs),which are the most common form of DNA variation,have great potential as a medical diagnostic tool.In this paper,the P-value is used as a SNP pre-selection criterion,and a wrapper algorithm with binary particle swarm optimization is used to find the rule for discriminating between affected and control subjects.We analyzed the association between combinations of SNPs and Graves' disease by investigating 108 SNPs in 384 cases and 652 controls.We evaluated our method by differentiating between cases and controls in a five-fold cross validation test,and it achieved a 72.9% prediction accuracy with a combination of 17 SNPs.The experimental results showed that SNPs,even those with a high P-value,have a greater effect on Graves' disease when acting in a combination.展开更多
基金supported in part by the National Science Foundation of China (61973247, 61673315, 62173268)the Key Research and Development Program of Shaanxi (2022GY-033)+2 种基金the Nationa Postdoctoral Innovative Talents Support Program of China (BX20200272)the Key Program of the National Natural Science Foundation of China (61833015)the Fundamental Research Funds for the Central Universities (xzy022021050)。
文摘The smart grid utilizes the demand side management technology to motivate energy users towards cutting demand during peak power consumption periods, which greatly improves the operation efficiency of the power grid. However, as the number of energy users participating in the smart grid continues to increase, the demand side management strategy of individual agent is greatly affected by the dynamic strategies of other agents. In addition, the existing demand side management methods, which need to obtain users’ power consumption information,seriously threaten the users’ privacy. To address the dynamic issue in the multi-microgrid demand side management model, a novel multi-agent reinforcement learning method based on centralized training and decentralized execution paradigm is presented to mitigate the damage of training performance caused by the instability of training experience. In order to protect users’ privacy, we design a neural network with fixed parameters as the encryptor to transform the users’ energy consumption information from low-dimensional to high-dimensional and theoretically prove that the proposed encryptor-based privacy preserving method will not affect the convergence property of the reinforcement learning algorithm. We verify the effectiveness of the proposed demand side management scheme with the real-world energy consumption data of Xi’an, Shaanxi, China. Simulation results show that the proposed method can effectively improve users’ satisfaction while reducing the bill payment compared with traditional reinforcement learning(RL) methods(i.e., deep Q learning(DQN), deep deterministic policy gradient(DDPG),QMIX and multi-agent deep deterministic policy gradient(MADDPG)). The results also demonstrate that the proposed privacy protection scheme can effectively protect users’ privacy while ensuring the performance of the algorithm.
基金supported by the National Natural Science Foundation of China(Nos.61402357,61272459,and 61402357)the China Postdoctoral Science Foundation(No.2015M570835)+2 种基金the Fundamental Research Funds for the Central Universities,Chinathe Program for New Century Excellent Talents in Universitythe CETC 54 Project(No.ITD-U14001/KX142600008)
文摘Controllers play a critical role in software-defined networking(SDN).However,existing singlecontroller SDN architectures are vulnerable to single-point failures,where a controller's capacity can be saturated by flooded flow requests.In addition,due to the complicated interactions between applications and controllers,the flow setup latency is relatively large.To address the above security and performance issues of current SDN controllers,we propose distributed rule store(DRS),a new multi-controller architecture for SDNs.In DRS,the controller caches the flow rules calculated by applications,and distributes these rules to multiple controller instances.Each controller instance holds only a subset of all rules,and periodically checks the consistency of flow rules with each other.Requests from switches are distributed among multiple controllers,in order to mitigate controller capacity saturation attack.At the same time,when rules at one controller are maliciously modified,they can be detected and recovered in time.We implement DRS based on Floodlight and evaluate it with extensive emulation.The results show that DRS can effectively maintain a consistently distributed rule store,and at the same time can achieve a shorter flow setup time and a higher processing throughput,compared with ONOS and Floodlight.
基金supported by the National Key R&D Program of China(Grant Nos.2018YFC0910400 and2017YFC0907500)the National Natural Science Foundation of China(Grant Nos.31671372 and 31701739)
文摘Every human being looks different in one way or the other.That’s the work of genetic variations,the ultimate driving force for evolution as well as the cause for many human diseases.Mapping human genetic variants reveals global genetic diversity,and pinpoints causal variants behind genetic disorders.
基金supported by the National Natural Science Foundation of China (Grant No. 60774086)the Research Fund for the Doctoral Program of Higher Education of China (Grant No. 20090201110027)
文摘Graves' disease,the production of thyroid-stimulating hormone receptor-stimulating antibodies leading to hyperthyroidism,is one of the most common forms of human autoimmune disease.It is widely agreed that complex diseases are not controlled simply by an individual gene or DNA variation but by their combination.Single nucleotide polymorphisms(SNPs),which are the most common form of DNA variation,have great potential as a medical diagnostic tool.In this paper,the P-value is used as a SNP pre-selection criterion,and a wrapper algorithm with binary particle swarm optimization is used to find the rule for discriminating between affected and control subjects.We analyzed the association between combinations of SNPs and Graves' disease by investigating 108 SNPs in 384 cases and 652 controls.We evaluated our method by differentiating between cases and controls in a five-fold cross validation test,and it achieved a 72.9% prediction accuracy with a combination of 17 SNPs.The experimental results showed that SNPs,even those with a high P-value,have a greater effect on Graves' disease when acting in a combination.