针对5G新空口-车联网(New Radio-Vehicle to Everything,NR-V2X)场景下车对基础设施(Vehicle to Infrastructure,V2I)和车对车(Vehicle to Vehicle,V2V)共享上行通信链路的频谱资源分配问题,提出了一种联邦-多智能体深度Q网络(Federated...针对5G新空口-车联网(New Radio-Vehicle to Everything,NR-V2X)场景下车对基础设施(Vehicle to Infrastructure,V2I)和车对车(Vehicle to Vehicle,V2V)共享上行通信链路的频谱资源分配问题,提出了一种联邦-多智能体深度Q网络(Federated Learning-Multi-Agent Deep Q Network,FL-MADQN)算法.该分布式算法中,每个车辆用户作为一个智能体,根据获取的本地信道状态信息,以网络信道容量最佳为目标函数,采用DQN算法训练学习本地网络模型.采用联邦学习加快以及稳定各智能体网络模型训练的收敛速度,即将各智能体的本地模型上传至基站进行聚合形成全局模型,再将全局模型下发至各智能体更新本地模型.仿真结果表明:与传统分布式多智能体DQN算法相比,所提出的方案具有更快的模型收敛速度,并且当车辆用户数增大时仍然保证V2V链路的通信效率以及V2I链路的信道容量.展开更多
In drilling field, stick-slip vibrations of the drill-string are the main reason for the failure of the drilling system. To suppress the undesired stick-slip vibrations, an observer-based state feedback control method...In drilling field, stick-slip vibrations of the drill-string are the main reason for the failure of the drilling system. To suppress the undesired stick-slip vibrations, an observer-based state feedback control method is proposed. The drilling system is described by a lumped parameter model including a Karnopp friction torque model. A state observer is designed to estimate the bit velocity in bottom hole and a state feedback controller is proposed to control the top drive velocity. By simulation, the performance of the control algorithm is demonstrated. Based on the control algorithm, a stick-slip vibration control system is developed. Test results show that the control system can effectively eliminate stick-slip vibrations of the drill-string and can be applied to the drilling field.展开更多
利用β-葡萄糖苷酶的转糖苷活性制备廉价葡萄糖-槐糖混合物(mixture of glucose and sophorose, MGD)作为诱导物被成功用于诱导里氏木霉(Trichoderma reesei)高效合成纤维素酶,但蛋白质组学显示相比于纤维素作为诱导物,膨胀素基因(Trsw...利用β-葡萄糖苷酶的转糖苷活性制备廉价葡萄糖-槐糖混合物(mixture of glucose and sophorose, MGD)作为诱导物被成功用于诱导里氏木霉(Trichoderma reesei)高效合成纤维素酶,但蛋白质组学显示相比于纤维素作为诱导物,膨胀素基因(Trswo1)显著性低表达导致纤维素酶水解效率低。因此利用组成型强启动子丙酮酸脱羧酶1(pyruvate decarboxylase1, PDC1)在T.reesei Rut C30中过表达内源Trswo1基因,2个T.reesei阳性转化子OEswo1-1和OEswo1-2的Trswo1基因转录水平分别提高了4.65倍和3.91倍,虽然T.reesei OEswo1-1和OEswo1-2在MGD诱导下合成的纤维素酶活性、内切纤维素酶活性和外切纤维素酶活性与野生菌酶活性均无显著性差异,但所产纤维素酶对玉米秸秆的水解效率分别提高了6.98%和13.93%。进一步通过X射线衍射(X-ray diffraction, XRD)和扫描电镜(scanning electron microscope, SEM)显示水解效率提高主要是由于木质纤维素结晶度的降低,同时纤维结构有不同程度的疏松和崩解。该研究成果对T.reesei利用可溶性诱导物合成纤维素酶系中纤维素降解辅助蛋白含量低的问题具有一定意义。展开更多
文摘针对5G新空口-车联网(New Radio-Vehicle to Everything,NR-V2X)场景下车对基础设施(Vehicle to Infrastructure,V2I)和车对车(Vehicle to Vehicle,V2V)共享上行通信链路的频谱资源分配问题,提出了一种联邦-多智能体深度Q网络(Federated Learning-Multi-Agent Deep Q Network,FL-MADQN)算法.该分布式算法中,每个车辆用户作为一个智能体,根据获取的本地信道状态信息,以网络信道容量最佳为目标函数,采用DQN算法训练学习本地网络模型.采用联邦学习加快以及稳定各智能体网络模型训练的收敛速度,即将各智能体的本地模型上传至基站进行聚合形成全局模型,再将全局模型下发至各智能体更新本地模型.仿真结果表明:与传统分布式多智能体DQN算法相比,所提出的方案具有更快的模型收敛速度,并且当车辆用户数增大时仍然保证V2V链路的通信效率以及V2I链路的信道容量.
文摘In drilling field, stick-slip vibrations of the drill-string are the main reason for the failure of the drilling system. To suppress the undesired stick-slip vibrations, an observer-based state feedback control method is proposed. The drilling system is described by a lumped parameter model including a Karnopp friction torque model. A state observer is designed to estimate the bit velocity in bottom hole and a state feedback controller is proposed to control the top drive velocity. By simulation, the performance of the control algorithm is demonstrated. Based on the control algorithm, a stick-slip vibration control system is developed. Test results show that the control system can effectively eliminate stick-slip vibrations of the drill-string and can be applied to the drilling field.
