短期预测在智能电网建设中扮演着重要角色,深刻影响电网发输变配用各个环节的智能化改造。短期预测一般基于系统实测数据,而传感器故障,数据传输错误等原因会导致数据质量下降,严重影响短期预测的精确性。为建立数据质量受损情况下的精...短期预测在智能电网建设中扮演着重要角色,深刻影响电网发输变配用各个环节的智能化改造。短期预测一般基于系统实测数据,而传感器故障,数据传输错误等原因会导致数据质量下降,严重影响短期预测的精确性。为建立数据质量受损情况下的精确短期预测模型,提出了结合数据预处理和双向长短期记忆(bi-directional long short-term memory,Bi-LSTM)的短期预测框架Bi-LSTM-DP(bi-directional long short-term memory data preprocessing)。在Bi-LSTM-DP中,采集的数据首先通过均值填补缺失值,进而基于Savitzky-Golay滤波器对数据降噪,最后采用Bi-LSTM提取时间序列的信息,实现短期预测。为了评估所提方法的性能,文中使用实测的公开数据集分别预测风电发电量和负荷需求,与其他参考方法对比表明了所述方法的有效性和鲁棒性。展开更多
At the 19th G20 Summit in Brazil in November 2024,China promoted the development of sustainable solutions to climate change,biodiversity loss,and environmental pollution.This continued the theme of the 2016 G20 Hangzh...At the 19th G20 Summit in Brazil in November 2024,China promoted the development of sustainable solutions to climate change,biodiversity loss,and environmental pollution.This continued the theme of the 2016 G20 Hangzhou Summit,at which China placed development at the center of the G20’s macroeconomic policy coordination for the first time,adopting the G20 Action Plan on the United Nations 2030 Agenda for Sustainable Development and the G20 Initiative on Supporting Industrialization in Africa and Least Developed Countries.In Brazil,China announced actions on advancing modernization in Africa over the next three years with a Chinese commitment of RMB360 billion yuan in financial support.In this article,we examine the potential role of geoscience research and practice in development,particularly in the sustainable use of natural resources,the prevention of climate change impacts,as well as mitigation of geo-hazards and their health implications,indicating the areas where China’s geoscience for Africa is strong and where it requires more effort.We find that although China is the world’s leading publisher of scientific papers,its contribution to geoscience in Africa(the globe’s fastest-growing economic area),as shown by bibliometric research,appears to be rather small and inconsistent with the research priorities of Africa.Amongst the priorities for geoscience research in Africa,which are not addressed substantially by the research conducted so far,are sustainable mineral and hydrocarbon development,hydrology and hydrogeology,climate change and resilience,natural hazards,medical geology,agrominerals,and geoscience education and training.A particular opportunity for African nations is the presence of critical minerals-minerals needed for the energy transition and for batteries for electric cars in particular.Africa is well-endowed with many of these critical materials,such as rare earth elements and platinum group metals.Several research groups stress the need for the agency on the part of African institutions to map out these valuable resources,understand their value and the economics and sustainability of their extraction,encourage local business,attract investment,and scrutinize proposals from potential international investors to get the best deals.A strong point of existing China-led geoscience development includes the Deep-time Digital Earth(DDE)program online computing platform and its artificial intelligence tool GeoGPT,which is being developed in partnership with Zhejiang Laboratory.These are being developed with strong China funding support for free and wide global access,with a particular focus on Africa.These advanced tools will help to place the agency of development squarely in the hands of African scientists and institutions.In summary,the following are recommended:(1)a more coordinated and strategic approach to China-led geoscience research in Africa;(2)an Africa-centered,geoscience funding initiative that concentrates on relevant topics to the continent such as critical minerals exploration and other geological resources,materials and processes and their health implications on the populations and ecosystems in general,as well as climate change and climate change resilience;and(3)continued support for China-led international initiatives that seek to increase the agency and capacity of Africa geoscience researchers,for example the Deep-time Digital Earth platform.展开更多
传统时序预测模型通常仅关注捕捉复杂时序中的趋势和模式,而忽略了变量间的相互作用,限制了该模型在复杂时序预测中应用.提出一种Dualformer双模型并联方案,该模型并联iTransformer(inverted transformer)和PatchTST(patch time series ...传统时序预测模型通常仅关注捕捉复杂时序中的趋势和模式,而忽略了变量间的相互作用,限制了该模型在复杂时序预测中应用.提出一种Dualformer双模型并联方案,该模型并联iTransformer(inverted transformer)和PatchTST(patch time series transformer),通过激活函数替代前馈神经网络,并通过多层感知机计算输出结果.Dualformer利用注意力机制同时捕捉复杂时序中的时间维度和变量维度信息,关注时间趋势与多变量交互.实验结果显示,Dualformer在复杂时序预测效果上显著优于对比模型iTransformer、PatchTST和DLinear(decomposition linear),在实际应用中可显著提高复杂时序预测的准确度,具有广泛应用前景.展开更多
现有研究在多QoS(quality of service)调度问题中,由于仅依赖即时奖励反馈机制,在资源受限的场景下处理时延敏感数据和具有连续传输需求的媒体数据时,存在可扩展性差和资源浪费的问题。为此,提出了一种基于奖励回溯的DQN(reward backtra...现有研究在多QoS(quality of service)调度问题中,由于仅依赖即时奖励反馈机制,在资源受限的场景下处理时延敏感数据和具有连续传输需求的媒体数据时,存在可扩展性差和资源浪费的问题。为此,提出了一种基于奖励回溯的DQN(reward backtracking based deep Q-network,RB-DQN)算法。该算法通过未来时刻的交互来回溯调整当前状态的策略评估,以更加有效地识别并解决因不合理调度策略导致的丢包。同时,设计了一种时延-吞吐均衡度量(latency throughput trade-off,LTT)指标,该指标综合考虑了时延敏感数据和媒体类型数据的业务需求,并可通过权重调整来突出不同的侧重点。大量仿真结果表明,与其他调度策略相比,所提算法能够有效降低时延敏感数据的延迟和抖动,同时确保媒体类型数据的流畅性与稳定性。展开更多
文摘短期预测在智能电网建设中扮演着重要角色,深刻影响电网发输变配用各个环节的智能化改造。短期预测一般基于系统实测数据,而传感器故障,数据传输错误等原因会导致数据质量下降,严重影响短期预测的精确性。为建立数据质量受损情况下的精确短期预测模型,提出了结合数据预处理和双向长短期记忆(bi-directional long short-term memory,Bi-LSTM)的短期预测框架Bi-LSTM-DP(bi-directional long short-term memory data preprocessing)。在Bi-LSTM-DP中,采集的数据首先通过均值填补缺失值,进而基于Savitzky-Golay滤波器对数据降噪,最后采用Bi-LSTM提取时间序列的信息,实现短期预测。为了评估所提方法的性能,文中使用实测的公开数据集分别预测风电发电量和负荷需求,与其他参考方法对比表明了所述方法的有效性和鲁棒性。
文摘At the 19th G20 Summit in Brazil in November 2024,China promoted the development of sustainable solutions to climate change,biodiversity loss,and environmental pollution.This continued the theme of the 2016 G20 Hangzhou Summit,at which China placed development at the center of the G20’s macroeconomic policy coordination for the first time,adopting the G20 Action Plan on the United Nations 2030 Agenda for Sustainable Development and the G20 Initiative on Supporting Industrialization in Africa and Least Developed Countries.In Brazil,China announced actions on advancing modernization in Africa over the next three years with a Chinese commitment of RMB360 billion yuan in financial support.In this article,we examine the potential role of geoscience research and practice in development,particularly in the sustainable use of natural resources,the prevention of climate change impacts,as well as mitigation of geo-hazards and their health implications,indicating the areas where China’s geoscience for Africa is strong and where it requires more effort.We find that although China is the world’s leading publisher of scientific papers,its contribution to geoscience in Africa(the globe’s fastest-growing economic area),as shown by bibliometric research,appears to be rather small and inconsistent with the research priorities of Africa.Amongst the priorities for geoscience research in Africa,which are not addressed substantially by the research conducted so far,are sustainable mineral and hydrocarbon development,hydrology and hydrogeology,climate change and resilience,natural hazards,medical geology,agrominerals,and geoscience education and training.A particular opportunity for African nations is the presence of critical minerals-minerals needed for the energy transition and for batteries for electric cars in particular.Africa is well-endowed with many of these critical materials,such as rare earth elements and platinum group metals.Several research groups stress the need for the agency on the part of African institutions to map out these valuable resources,understand their value and the economics and sustainability of their extraction,encourage local business,attract investment,and scrutinize proposals from potential international investors to get the best deals.A strong point of existing China-led geoscience development includes the Deep-time Digital Earth(DDE)program online computing platform and its artificial intelligence tool GeoGPT,which is being developed in partnership with Zhejiang Laboratory.These are being developed with strong China funding support for free and wide global access,with a particular focus on Africa.These advanced tools will help to place the agency of development squarely in the hands of African scientists and institutions.In summary,the following are recommended:(1)a more coordinated and strategic approach to China-led geoscience research in Africa;(2)an Africa-centered,geoscience funding initiative that concentrates on relevant topics to the continent such as critical minerals exploration and other geological resources,materials and processes and their health implications on the populations and ecosystems in general,as well as climate change and climate change resilience;and(3)continued support for China-led international initiatives that seek to increase the agency and capacity of Africa geoscience researchers,for example the Deep-time Digital Earth platform.
文摘传统时序预测模型通常仅关注捕捉复杂时序中的趋势和模式,而忽略了变量间的相互作用,限制了该模型在复杂时序预测中应用.提出一种Dualformer双模型并联方案,该模型并联iTransformer(inverted transformer)和PatchTST(patch time series transformer),通过激活函数替代前馈神经网络,并通过多层感知机计算输出结果.Dualformer利用注意力机制同时捕捉复杂时序中的时间维度和变量维度信息,关注时间趋势与多变量交互.实验结果显示,Dualformer在复杂时序预测效果上显著优于对比模型iTransformer、PatchTST和DLinear(decomposition linear),在实际应用中可显著提高复杂时序预测的准确度,具有广泛应用前景.
文摘现有研究在多QoS(quality of service)调度问题中,由于仅依赖即时奖励反馈机制,在资源受限的场景下处理时延敏感数据和具有连续传输需求的媒体数据时,存在可扩展性差和资源浪费的问题。为此,提出了一种基于奖励回溯的DQN(reward backtracking based deep Q-network,RB-DQN)算法。该算法通过未来时刻的交互来回溯调整当前状态的策略评估,以更加有效地识别并解决因不合理调度策略导致的丢包。同时,设计了一种时延-吞吐均衡度量(latency throughput trade-off,LTT)指标,该指标综合考虑了时延敏感数据和媒体类型数据的业务需求,并可通过权重调整来突出不同的侧重点。大量仿真结果表明,与其他调度策略相比,所提算法能够有效降低时延敏感数据的延迟和抖动,同时确保媒体类型数据的流畅性与稳定性。
