检索结果-内蒙古大学图书馆

Energy-Efficient stochastic Computing (SC) Neural Networks for Internet of Things Devices With Layer-Wise Adjustable Sequence Length (ASL)

引用

IEEE Internet of Things Journal 2025年

作者： Wang, Ziheng Reviriego, Pedro Niknia, Farzad Gao, Zhen Conde, Javier Liu, Shanshan Lombardi, Fabrizio Northeastern University Department of Electrical and Computer Engineering MA02115 United States Universidad Politécnica de Madrid ETSI de Telecomunicación Madrid28040 Spain Tianjin University School of Electrical and Information Engineering Tianjin300072 China University of Electronic Science and Technology of China School of Information and Communication Engineering Chengdu611731 China

stochastic computing (SC) has emerged as an efficient low-power alternative for deploying neural networks (NNs) in resource-limited scenarios, such as the Internet of Things (IoT). By encoding values as serial bitstreams, SC significantly reduces energy dissipation compared to conventional floating-point (FP) designs;however, further improvement of layer-wise mixed-precision implementation for SC remains unexplored. This paper introduces Adjustable Sequence Length (ASL), a novel scheme that applies mixedprecision concepts specifically to SC NNs. By introducing an operator-norm – based theoretical model, this paper shows that truncation noise can cumulatively propagate through the layers by the estimated amplification factors. An extended sensitivity analysis is presented, using Random Forest (RF) regression to evaluate multi-layer truncation effects and validate the alignment of theoretical predictions with practical network behaviors. To accommodate different application scenarios, this paper proposes two truncation strategies (coarse-grained and fine-grained), which apply diverse sequence length configurations at each layer. Evaluations on a pipelined SC MLP synthesized at 32 nm demonstrate that ASL can reduce energy and latency overheads by up to over 60% with negligible accuracy loss. It confirms the feasibility of the ASL scheme for IoT applications and highlights the distinct advantages of mixed-precision truncation in SC designs. © 2014 IEEE.

关键词： stochastic programming

来源：评论

学校读者我要写书评

暂无评论

Capacitated hub location routing problem with time windows and stochastic demands for the design of intra-city express systems

引用

European Journal of Operational Research 2025年第2期326卷 255-269页

作者： Wu, Yuehui Fang, Hui Qureshi, Ali Gul Yamada, Tadashi College of Transportation Engineering Dalian Maritime University Liaoning Dalian116026 China Department of Urban Management of Engineering Kyoto University Kyoto 615-8246 Japan of Management Kyoto University Yoshida Honmachi Sakyo-ku Kyoto Kyoto-shi606-8501 Japan

This work focuses on planning an intra-city express system in a practical environment. Various operation characteristics, such as vehicle capacity, hub capacity, time windows, and stochastic demands, have been considered. Therefore, we introduce a capacitated hub location routing problem with time windows and stochastic demand and formulate it using a multi-stage recourse model. In this model, long-term decisions (hub location and client-to-hub allocation) are made first, and short-term decisions (vehicle routing) are determined after revealing stochastic variables. To solve the problem, we propose a hybrid stochastic variable neighbourhood search (HSVNS) algorithm, which integrates an adaptive large neighbourhood search (ALNS) algorithm within a stochastic variable neighbourhood search (SVNS) framework. Numerical experiments and case studies indicate that the HSVNS algorithm can provide high-quality solutions within a reasonable computation time for instances with up to 70 clients and that considering stochastic factors can efficiently reduce operation costs, especially for instances with tight vehicle capacity and loose time windows. © 2025 Elsevier B.V.

关键词： stochastic programming

来源：评论

学校读者我要写书评

暂无评论

Dynamic capacitated facility location problem in mobile renewable energy charging stations under sustainability consideration

引用

SUSTAINABLE COMPUTING-INFORMATICS & SYSTEMS 2024年 41卷

作者： Ala, Ali Deveci, Muhammet Bani, Erfan Amani Sadeghi, Amir Hossein Shanghai Jiao Tong Univ Dept Ind Engn & Management Sch Mech Engn Shanghai 200240 Peoples R China Natl Def Univ Turkish Naval Acad Dept Ind Engn TR-34940 Istanbul Turkiye Lebanese Amer Univ Dept Elect & Comp Engn Byblos Lebanon Imperial Coll London Royal Sch Mines South Kensington Campus London SW7 2AZ England Sharif Univ Technol Dept Ind Engn Tehran Iran North Carolina State Univ Dept Ind & Syst Engn Raleigh NC USA

The deployment of mobile renewable energy charging stations plays a crucial role in facilitating the overall adoption of electric vehicles and reducing reliance on fossil fuels. This study addresses the dynamic capacitated facility location problem in mobile charging stations from a sustainability perspective. This paper proposes Twostage stochastic programming with recourse that performs well for this application, and the location of the mobile renewable energy charging station (MRECS) management addresses the complex dynamics of reusable items. To solve this problem, we suggested dealing with differential evolutionary (DE) and DE Q-learning (DEQL) algorithms, as two novel optimization and reinforcement learning approaches, are presented as solution approaches to validate their performance. Evaluation of the outcomes reveals a considerable disparity between the algorithms, and DEQL performs better in solving the presented problem. In addition, DEQL could minimize the total operation cost and carbon emission by 7% and 20%, respectively. In contrast, the DE could decrease carbon emissions and total operation costs by 5% and 2.5%, respectively.

关键词： Renewable energy stochastic programming Optimization algorithm Reinforcement learning Facility location Sustainability

来源：评论

学校读者我要写书评

暂无评论

Optimal cruise airspeed selection and RTA adjustment in the presence of wind uncertainty

引用

TRANSPORTATION RESEARCH PART C-EMERGING TECHNOLOGIES 2024年 162卷

作者： Jung, Hyunho Clarke, John-Paul Univ Texas Austin Austin TX 78712 USA Univ Texas Austin Walker Dept Mech Engn Austin TX 78712 USA Univ Texas Austin Dept Aerosp Engn & Engn Mech Austin TX 78712 USA

Flight planning under windy conditions has been a critical problem in air transportation. Wind information is primarily obtained through hourly updated forecasts available prior to the flight. Although flight plans are made based on the forecast, controlling the time of arrival at the destination airport can still be difficult. Furthermore, the imprecision and uncertainty in weather forecasts present additional challenges to flight operations. In view of this, we propose an approach for selecting the optimal cruise airspeed under wind uncertainty to save fuel and maintain the Required Time of Arrival (RTA). The entire flight is first divided into a certain number of segments, and the correlations between wind forecasts and actual observations are utilized by merging stochastic programming (SP) and Receding Horizon Control (RHC) frameworks to determine the optimum airspeed. Additionally, we propose a method to reduce fuel consumption across multiple flights by sharing information between aircraft and the air traffic center. This information is then used to determine RTAs and adjust them to achieve fuel efficiency. Through numerical studies, we obtained satisfactory results indicating that the proposed method can lead to improved performance in fuel saving and RTA for all flights in the experiment.

关键词： Air traffic management stochastic programming Receding Horizon Control Optimal airspeed search

来源：评论

学校读者我要写书评

暂无评论

A chance-constrained stochastic chiller sequencing strategy considering life-expectancy of chiller plant

引用

ENERGY 2024年 290卷

作者： Kumar, Devesh Pindoriya, Naran M. Indian Inst Technol Gandhinagar Elect Engn Gandhinagar 382355 Gujarat India

This article focuses on addressing the chiller sequencing problem of chiller plants by establishing a comprehensive energy management framework. The contribution of this work is threefold. Firstly, a distributive modeling architecture is presented that establishes five concurrent models as input to the framework. A synergy among these models is exploited to formulate a chance-constrained stochastic chiller sequencing problem. Secondly, a quantified life expectancy model of the chiller plant is introduced and a case is made for why it is influential in delivering industrially applicable solutions. The model attempts to strike a balance between economic optimality and improved reliability. Thirdly, a chiller data pre-processing protocol incubating two heuristic algorithms is proposed to address measurement uncertainties of the chiller plant state variables. Furthermore, this work develops a robust ensemble model to accurately forecast the cooling load and embeds a chain of reformulations to improve the global solution's optimality. The developed framework is realized with the plant at the Indian Institute of Technology Gandhinagar. The results confirm that the proposed framework leads to a significant amount of power savings. In comparison to conventional scheduling, the chiller plant power consumption can be reduced by up to 6.2 %, thereby illustrating its efficacy.

关键词： Chiller plant Chiller sequencing Energy management Operational optimization stochastic programming

来源：评论

学校读者我要写书评

暂无评论

A smart building energy management incorporating clustering-based tariffs in the presence of domestic solar energy, battery, and electric vehicle

引用

SOLAR ENERGY 2024年 279卷

作者： Alilou, Masoud Mohammadi-ivatloo, Behnam Mohammadpourfard, Mousa Univ Tabriz Fac Elect & Comp Engn Tabriz Iran LUT Univ Sch Energy Syst Dept Elect Engn Lappeenranta Finland Univ Tabriz Fac Chem & Petr Engn Tabriz Iran Izmir Inst Technol Dept Energy Syst Engn Izmir Turkiye

Smart buildings play a crucial role in optimizing energy management within the power network. As end-users of the power network, they have the ability to not only reduce economic costs for householders but also modify the technical indices of the power network. To promote efficient device management in smart homes (SH), demand response programs are recommended for consumers. This research investigates the application of clusteringbased electricity pricing strategy aimed at effectively managing the energy devices of a residential smart home. The utilized method categorizes the electricity tariff into five rates according to the clustering of the realtime pricing program. Ward's clustering method is utilized to cluster and determine new electricity tariffs. The primary goal of the energy management program is to minimize the building's energy cost, which is accomplished through the utilization of the multi-verse optimizer. The smart home consists of essential and manageable appliances, a photovoltaic panel (PV), a sodium-sulfur (NaS) battery, and an electric vehicle (EV). The initial parameters of the PV and EV are modeled stochastically by their probability distribution functions and calculated using the Latin hypercube sampling algorithm. The smart building's performance is assessed by taking into account various demand response programs. The numerical results present that the application of the clusteringbased management method has resulted in a significant reduction of 23-43 % in the electricity cost of smart homes. Additionally, the smart home exhibits a more linear consumption pattern when considering the electricity tariffs based on the clustering approach.

关键词： Clustering-based electricity tariff Smart home Solar energy stochastic programming Demand response program Electric vehicle

来源：评论

学校读者我要写书评

暂无评论

Residential demand response in the European power system: No significant impact on capacity expansion and cost savings

引用

SUSTAINABLE ENERGY GRIDS & NETWORKS 2024年 38卷

作者： Barani, Mostafa Backe, Stian O'Reilly, Ryan del Granado, Pedro Crespo Norwegian Univ Sci & Technol NTNU Dept Ind Econ & Technol Management Trondheim Norway SINTEF Energy Res Energy Syst Trondheim Norway Xcel Energy Fleet Emiss Vehicles 0 Ft Collins CO USA

Direct Load Control (DLC) is a demand response strategy in which customers receive compensation from utilities in return for permitting them to regulate the operation of specific equipment. This paper analyzes the impacts of DLC programs on the transition to renewable energy within the European electricity system towards 2060. The study quantifies the achievable hourly potential for DLC across Europe in the residential sector. By implementing and developing a DLC module within the stochastic capacity expansion model EMPIRE, we investigate how costs, long-term investments, and long-term marginal prices are affected by residential DLC participation rates. The research utilizes a comprehensive DLC dataset, including ten appliances such as electric vehicles, heat pumps, refrigeration, and others. This dataset serves as the basis for creating four storylines to investigate the integration of these programs into the European electricity system. The results indicate that residential DLC programs have some impact on grid -battery deployment, PV plant penetration, and electricity prices. In the best -case scenario, involving ambitious participation of residential loads in DLC programs without compensation, cost savings are about 1% versus not introducing DLC. The findings contribute to understanding the value of demand response programs in Europe, indicating that the savings they bring might not be sufficient to provide enough incentives or compensation for widespread participation in such programs. That is, from a long-term investment or capacity expansion perspective, it may not be worthwhile to soley include residential demand response in the planning of the electricity system.

关键词： Demand response program Direct load control European energy transition Power system investment Flexibility stochastic programming

来源：评论

学校读者我要写书评

暂无评论

Reinforcement learning approaches for the stochastic discrete lot-sizing problem on parallel machines

引用

EXPERT SYSTEMS WITH APPLICATIONS 2024年 246卷

作者： Felizardo, Leonardo Kanashiro Fadda, Edoardo Del-Moral-Hernandez, Emilio Brandimarte, Paolo Univ Sao Paulo Escola Politecn Ave Prof Luciano Gualberto 380 Sao Paulo Brazil Politecn Torino Dipartimento Sci Matemat Corso Duca Abruzzi 24 I-10129 Turin Italy

This paper addresses the stochastic discrete lot-sizing problem on parallel machines, which is a computationally challenging problem also for relatively small instances. We propose two heuristics to deal with it by leveraging reinforcement learning. In particular, we propose a technique based on approximate value iteration around post-decision state variables and one based on multi-agent reinforcement learning. We compare these two approaches with other reinforcement learning methods and more classical solution techniques, showing their effectiveness in addressing realistic size instances.

关键词： Dynamic programming stochastic programming Multi-agent systems Machine learning Reinforcement Learning

来源：评论

学校读者我要写书评

暂无评论

Resource Allocation in Quantum Key Distribution (QKD) for Space-Air-Ground Integrated Networks 27

Resource Allocation in Quantum Key Distribution (QKD) for Sp...

引用

27th International Workshop on Computer Aided Modeling and Design of Communication Links and Networks (CAMAD)

作者： Kaewpuang, Rakpong Xu, Minrui Niyato, Dusit Yu, Han Xiong, Zehui Nanyang Technol Univ NTU Sch Comp Sci & Engn Singapore Singapore Singapore Univ Technol & Design SUTD Pillar Informat Syst Technol & Design Singapore Singapore

ISBN: (数字)9781665461290

ISBN: (纸本)9781665461290

Space-air-ground integrated networks (SAGIN) are one of the most promising advanced paradigms in the sixth generation (6G) communication. SAGIN can support high data rates, low latency, and seamless network coverage for interconnected applications and services. However, communications in SAGIN are facing tremendous security threats from the everincreasing capacity of quantum computers. Fortunately, quantum key distribution (QKD) for establishing secure communications in SAGIN, i.e., QKD over SAGIN, can provide informationtheoretic security. To minimize the QKD deployment cost in SAGIN with heterogeneous nodes, in this paper, we propose a resource allocation scheme for QKD over SAGIN using stochastic programming. The proposed scheme is formulated via two-stage stochastic programming (SP), while considering uncertainties such as security requirements and weather conditions. Under extensive experiments, the results clearly show that the proposed scheme can achieve the optimal deployment cost under various security requirements and unpredictable weather conditions.

关键词： Quantum key distribution space-air-ground integrated networks resource allocation stochastic programming

来源：评论

学校读者我要写书评

暂无评论

Bidding and Scheduling in Energy Markets: Which Probabilistic Forecast Do We Need? 17

Bidding and Scheduling in Energy Markets: Which Probabilisti...

引用

17th International Conference on Probabilistic Methods Applied to Power Systems (PMAPS)

作者： Beykirch, Mario Janke, Tim Steinke, Florian Tech Univ Darmstadt Energy Informat Networks & Syst Darmstadt Germany

ISBN: (纸本)9781665412117

Probabilistic forecasting in combination with stochastic programming is a key tool for handling the growing uncertainties in future energy systems. Derived from a general stochastic programming formulation for the optimal scheduling and bidding in energy markets we examine several common special instances containing uncertain loads, energy prices, and variable renewable energies. We analyze for each setup whether only an expected value forecast, marginal or bivariate predictive distributions, or the full joint predictive distribution is required. For market schedule optimization, we find that expected price forecasts are sufficient in almost all cases, while the marginal distributions of renewable energy production and demand are often required. For bidding curve optimization, pairwise or full joint distributions are necessary except for specific cases. This work helps practitioners choose the simplest type of forecast that can still achieve the best theoretically possible result for their problem and researchers to focus on the most relevant instances.

关键词： stochastic programming probabilistic forecasting energy forecasting energy scheduling

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：