Based on the theory of complex adaptive system(CAS),the optimal allocation model of water resources in sewage irrigation areas was established,which provided new ideas and application value for the rational utilizatio...Based on the theory of complex adaptive system(CAS),the optimal allocation model of water resources in sewage irrigation areas was established,which provided new ideas and application value for the rational utilization of agricultural production and waste water resources.The results demonstrated that the difference of crop energy capture mainly depended on the development stage.Waste water with a certain concentration was able to promote crop growth,while excessive concentration inhibited crop growth.The correlation between water absorption rate and leaf area index was close(R=0.9498,p<0.01).The amount of bad seeds increased at a speed of 34.7·d^-1,when system irrigated randomly in the seedling stage,while it tended to remain stable at a speed of 0.3·d^-1 after plants entering the mature stage which impacted the total yields of crops.展开更多
在相控阵天线跳波束对大范围潜在区域进行捷变覆盖的问题中,当前资源分配方法仅针对单一资源分配优化达到局部最优并未有统筹考虑,从空域、频域、时域、调制编码域等多个维度全面建立了相控阵跳波束下的资源联合分配优化模型,给出了一...在相控阵天线跳波束对大范围潜在区域进行捷变覆盖的问题中,当前资源分配方法仅针对单一资源分配优化达到局部最优并未有统筹考虑,从空域、频域、时域、调制编码域等多个维度全面建立了相控阵跳波束下的资源联合分配优化模型,给出了一种基于遗传算法和动态规划的模型求解方法。仿真结果表明,考虑了多维资源进行联合分配的方法,可有效降低卫星通信网络中各终端的缓存队列长度,从而提高用户服务质量(Quality of Service,QoS)及网络吞吐量。展开更多
在资源受限的水声网络中,使用软频率复用技术和自适应资源分配技术可以提高网络容量和能量效率。然而,水声信道的长传播时延和时变特性导致用于自适应技术的反馈信道状态信息(Channel State Information, CSI)是时变且过时的。非理想的...在资源受限的水声网络中,使用软频率复用技术和自适应资源分配技术可以提高网络容量和能量效率。然而,水声信道的长传播时延和时变特性导致用于自适应技术的反馈信道状态信息(Channel State Information, CSI)是时变且过时的。非理想的反馈CSI会降低自适应系统的性能。针对该问题,提出了一种基于多智能体深度Q网络的资源分配(Multi-agent Deep Q Network Based Resource Allocation, MADQN-RA)方法。该方法将水声软频率复用网络视为多智能体系统,并使用过时的反馈CSI序列作为系统状态。通过建立有效的奖励表达式,智能体可以跟踪时变时延水声信道的变化特性并做出相应的资源分配决策。为了进一步提高智能体的决策准确度,同时避免状态空间维度增大时的部分学习成本,结合动态状态长度方法改进了MADQN-RA。仿真结果表明,所提方法实现的系统性能优于基于其他学习的方法和基于信道预测的方法,且更接近理论最优值。展开更多
基金Supported by the Science and Technology Research Project of the Ministry of Education(14YJCZH017)the Major State Basic Research Development Program of China(973 Program)(2017YFC0404503)+1 种基金Key Cultivation Project of Lingnan Normal University in 2019(LZ1903)Lingnan Normal University Special Talent Program(ZL2007)
文摘Based on the theory of complex adaptive system(CAS),the optimal allocation model of water resources in sewage irrigation areas was established,which provided new ideas and application value for the rational utilization of agricultural production and waste water resources.The results demonstrated that the difference of crop energy capture mainly depended on the development stage.Waste water with a certain concentration was able to promote crop growth,while excessive concentration inhibited crop growth.The correlation between water absorption rate and leaf area index was close(R=0.9498,p<0.01).The amount of bad seeds increased at a speed of 34.7·d^-1,when system irrigated randomly in the seedling stage,while it tended to remain stable at a speed of 0.3·d^-1 after plants entering the mature stage which impacted the total yields of crops.
文摘在相控阵天线跳波束对大范围潜在区域进行捷变覆盖的问题中,当前资源分配方法仅针对单一资源分配优化达到局部最优并未有统筹考虑,从空域、频域、时域、调制编码域等多个维度全面建立了相控阵跳波束下的资源联合分配优化模型,给出了一种基于遗传算法和动态规划的模型求解方法。仿真结果表明,考虑了多维资源进行联合分配的方法,可有效降低卫星通信网络中各终端的缓存队列长度,从而提高用户服务质量(Quality of Service,QoS)及网络吞吐量。
文摘在资源受限的水声网络中,使用软频率复用技术和自适应资源分配技术可以提高网络容量和能量效率。然而,水声信道的长传播时延和时变特性导致用于自适应技术的反馈信道状态信息(Channel State Information, CSI)是时变且过时的。非理想的反馈CSI会降低自适应系统的性能。针对该问题,提出了一种基于多智能体深度Q网络的资源分配(Multi-agent Deep Q Network Based Resource Allocation, MADQN-RA)方法。该方法将水声软频率复用网络视为多智能体系统,并使用过时的反馈CSI序列作为系统状态。通过建立有效的奖励表达式,智能体可以跟踪时变时延水声信道的变化特性并做出相应的资源分配决策。为了进一步提高智能体的决策准确度,同时避免状态空间维度增大时的部分学习成本,结合动态状态长度方法改进了MADQN-RA。仿真结果表明,所提方法实现的系统性能优于基于其他学习的方法和基于信道预测的方法,且更接近理论最优值。
