针对航天器与非合作目标追逃博弈的生存型微分对策拦截问题,基于强化学习研究了追逃博弈策略,提出了自适应增强随机搜索(adaptive-augmented random search,A-ARS)算法。针对序贯决策的稀疏奖励难题,设计了基于策略参数空间扰动的探索方...针对航天器与非合作目标追逃博弈的生存型微分对策拦截问题,基于强化学习研究了追逃博弈策略,提出了自适应增强随机搜索(adaptive-augmented random search,A-ARS)算法。针对序贯决策的稀疏奖励难题,设计了基于策略参数空间扰动的探索方法,加快策略收敛速度;针对可能过早陷入局部最优问题设计了新颖度函数并引导策略更新,可提升数据利用效率;通过数值仿真验证并与增强随机搜索(augmented random search,ARS)、近端策略优化算法(proximal policy optimization,PPO)以及深度确定性策略梯度下降算法(deep deterministic policy gradient,DDPG)进行对比,验证了此方法的有效性和先进性。展开更多
负载预测的精度是影响云平台弹性资源管理的主要因素之一。而云平台中存在着大量的短任务负载序列,其历史信息不足和不平滑的特性导致难以选择合适的模型进行精准预测。对此提出了一种领域对抗自适应的短任务负载预测模型。该模型采用...负载预测的精度是影响云平台弹性资源管理的主要因素之一。而云平台中存在着大量的短任务负载序列,其历史信息不足和不平滑的特性导致难以选择合适的模型进行精准预测。对此提出了一种领域对抗自适应的短任务负载预测模型。该模型采用奇异谱分析(singular spectrum analysis,SSA)对样本进行平滑处理;联合第四版本的Mueen相似度搜索算法(the fourth version of Mueen’s algorithm for similarity search,MASS_V4)与时间特征进行域间相似性计算,获得合适的源域数据来辅助迁移预测;将门控循环单元(gated recurrent unit,GRU)作为基准器构建网络,并利用Y差异定义新的损失函数,通过对抗过程建立出表征能力强的短任务负载预测模型。将所提方法在两个真实的云平台数据集上与其他常用的云负载预测算法对比,均表现出较高的预测精度。展开更多
To understand the pollution characteristics and distributions of microplastics in the Dagu River,an important inland river that flows to Jiaozhou Bay,the abundance,particle size,shape,color,and polymer type of micropl...To understand the pollution characteristics and distributions of microplastics in the Dagu River,an important inland river that flows to Jiaozhou Bay,the abundance,particle size,shape,color,and polymer type of microplastics in the surface waters and floodplain sediments were investigated.The results showed that the abundance of microplastics in the surface waters varied between 0.68 and 3.96 items L^(−1).The abundance of microplastics in the floodplain sediments varied between 115.5 and 495 items kg^(−1).The distribution of microplastics in the river has obvious characteristics:1)The abundance of microplastics downstream was significantly higher than that upstream.2)The abundance of microplastics decreased significantly in the estuary.3)The abundance of microplastics was higher in densely populated areas.4)Factors such as effluent discharge from sewage treatment plants and inflow of tributaries affect the abundance of microplastics in rivers.The main polymer types of microplastics were polypropylene(PP),polyethylene(PE),and polyvinyl chloride(PVC).Both the abundance and polymer types of microplastics show a positive correlation between the surface waters and floodplain sediment samples,indicating that the microplastics in surface waters and floodplain sediments come from the same source.The present study demonstrates that the abundance,distribution,and sources of microplastics might be affected by geology,hydrogeology,land use types,and anthropogenic activities.This study provides basic data for the study of microplastics in rivers and provides a reference for the establishment of microplastic control measures.展开更多
目的分析5种识别α-突触核蛋白(α-synuclein,α-syn)N端结构域的单克隆抗体1C16、2B8、2P21、3O18和1J6的特异性,为帕金森病(Parkinson s disease,PD)的早期诊断和免疫治疗提供抗体支持。方法体外蛋白重组技术获得人源α-syn蛋白(human...目的分析5种识别α-突触核蛋白(α-synuclein,α-syn)N端结构域的单克隆抗体1C16、2B8、2P21、3O18和1J6的特异性,为帕金森病(Parkinson s disease,PD)的早期诊断和免疫治疗提供抗体支持。方法体外蛋白重组技术获得人源α-syn蛋白(human-α-syn,h-α-syn)、鼠源α-syn蛋白(mouse-α-syn,m-α-syn)、β-syn蛋白、N端人源α-syn蛋白(α-syn/N)和去N端人源α-syn蛋白(α-syn/ΔN)。斑点印迹法鉴定5种单克隆抗体的特异性识别结构域,使用蛋白质印迹技术检测其对变性后的纯蛋白和鼠脑组织的识别情况。结果抗体1C16可以识别非变性的h-α-syn纯蛋白和小鼠脑组织中的变性的h-α-syn蛋白,不识别m-α-syn和β-syn。抗体1J6仅识别非变性的h-α-syn及α-syn/N纯蛋白,不识别变性后的全长α-syn纯蛋白和小鼠脑组织中的变性h-α-syn蛋白。结论筛选出的2种单克隆抗体具有特异性,可为本实验下一步生物标志物的酶联免疫吸附法测定检测和针对α-syn的免疫治疗提供基础。展开更多
文摘针对航天器与非合作目标追逃博弈的生存型微分对策拦截问题,基于强化学习研究了追逃博弈策略,提出了自适应增强随机搜索(adaptive-augmented random search,A-ARS)算法。针对序贯决策的稀疏奖励难题,设计了基于策略参数空间扰动的探索方法,加快策略收敛速度;针对可能过早陷入局部最优问题设计了新颖度函数并引导策略更新,可提升数据利用效率;通过数值仿真验证并与增强随机搜索(augmented random search,ARS)、近端策略优化算法(proximal policy optimization,PPO)以及深度确定性策略梯度下降算法(deep deterministic policy gradient,DDPG)进行对比,验证了此方法的有效性和先进性。
文摘负载预测的精度是影响云平台弹性资源管理的主要因素之一。而云平台中存在着大量的短任务负载序列,其历史信息不足和不平滑的特性导致难以选择合适的模型进行精准预测。对此提出了一种领域对抗自适应的短任务负载预测模型。该模型采用奇异谱分析(singular spectrum analysis,SSA)对样本进行平滑处理;联合第四版本的Mueen相似度搜索算法(the fourth version of Mueen’s algorithm for similarity search,MASS_V4)与时间特征进行域间相似性计算,获得合适的源域数据来辅助迁移预测;将门控循环单元(gated recurrent unit,GRU)作为基准器构建网络,并利用Y差异定义新的损失函数,通过对抗过程建立出表征能力强的短任务负载预测模型。将所提方法在两个真实的云平台数据集上与其他常用的云负载预测算法对比,均表现出较高的预测精度。
基金financially supported by the National Major Special Project for Water Pollution Treatment and Control ‘Technology Integration of Watershed Ecological Function Zoning Management’ Project (No. 2017ZX073 01-001)the National Key Research and Development Program of China (No. 2018YFC1407601)。
文摘To understand the pollution characteristics and distributions of microplastics in the Dagu River,an important inland river that flows to Jiaozhou Bay,the abundance,particle size,shape,color,and polymer type of microplastics in the surface waters and floodplain sediments were investigated.The results showed that the abundance of microplastics in the surface waters varied between 0.68 and 3.96 items L^(−1).The abundance of microplastics in the floodplain sediments varied between 115.5 and 495 items kg^(−1).The distribution of microplastics in the river has obvious characteristics:1)The abundance of microplastics downstream was significantly higher than that upstream.2)The abundance of microplastics decreased significantly in the estuary.3)The abundance of microplastics was higher in densely populated areas.4)Factors such as effluent discharge from sewage treatment plants and inflow of tributaries affect the abundance of microplastics in rivers.The main polymer types of microplastics were polypropylene(PP),polyethylene(PE),and polyvinyl chloride(PVC).Both the abundance and polymer types of microplastics show a positive correlation between the surface waters and floodplain sediment samples,indicating that the microplastics in surface waters and floodplain sediments come from the same source.The present study demonstrates that the abundance,distribution,and sources of microplastics might be affected by geology,hydrogeology,land use types,and anthropogenic activities.This study provides basic data for the study of microplastics in rivers and provides a reference for the establishment of microplastic control measures.
文摘目的分析5种识别α-突触核蛋白(α-synuclein,α-syn)N端结构域的单克隆抗体1C16、2B8、2P21、3O18和1J6的特异性,为帕金森病(Parkinson s disease,PD)的早期诊断和免疫治疗提供抗体支持。方法体外蛋白重组技术获得人源α-syn蛋白(human-α-syn,h-α-syn)、鼠源α-syn蛋白(mouse-α-syn,m-α-syn)、β-syn蛋白、N端人源α-syn蛋白(α-syn/N)和去N端人源α-syn蛋白(α-syn/ΔN)。斑点印迹法鉴定5种单克隆抗体的特异性识别结构域,使用蛋白质印迹技术检测其对变性后的纯蛋白和鼠脑组织的识别情况。结果抗体1C16可以识别非变性的h-α-syn纯蛋白和小鼠脑组织中的变性的h-α-syn蛋白,不识别m-α-syn和β-syn。抗体1J6仅识别非变性的h-α-syn及α-syn/N纯蛋白,不识别变性后的全长α-syn纯蛋白和小鼠脑组织中的变性h-α-syn蛋白。结论筛选出的2种单克隆抗体具有特异性,可为本实验下一步生物标志物的酶联免疫吸附法测定检测和针对α-syn的免疫治疗提供基础。
