检索结果-内蒙古大学图书馆

37th Conference on Neural Information Processing Systems (NeurIPS)

作者： Shen, Zhenqian Yang, Hansi Li, Yong Kwok, James Yao, Quanming Tsinghua Univ Dept Elect Engn Beijing Peoples R China Hong Kong Univ Sci & Technol Dept Comp Sci & Engn Hong Kong Peoples R China

ISBN: (纸本)9781713899921

As hyper-parameters are ubiquitous and can significantly affect the model performance, hyper-parameter optimization is extremely important in machine learning. In this paper, we consider a sub-class of hyper-parameter optimization problems, where the hyper-gradients are not available. Such problems frequently appear when the performance metric is non-differentiable or the hyper-parameter is not continuous. However, existing algorithms, like Bayesian optimization and reinforcement learning, often get trapped in local optimals with poor performance. To address the above limitations, we propose to use cubic regularization to accelerate convergence and avoid saddle points. First, we adopt stochastic relaxation, which allows obtaining gradient and Hessian information without hyper-gradients. Then, we exploit the rich curvature information by cubic regularization. Theoretically, we prove that the proposed method can converge to approximate second-order stationary points, and the convergence is also guaranteed when the lower-level problem is inexactly solved. Experiments on synthetic and real-world data demonstrate the effectiveness of our proposed method.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

An Efficient Antenna optimization Method Based on Variable Population Size 12

An Efficient Antenna Optimization Method Based on Variable P...

引用

12th IEEE Asia-Pacific Conference on Antennas and Propagation, APCAP 2024

作者： Peng, Fengling Chen, Xing Sichuan University Sichuan Chengdu610065 China

ISBN: (纸本)9798350351019

To reduce the computational cost and time associated with automatic antenna optimization, this paper proposes a differential evolution algorithm with a variable population mechanism. Firstly, the paper introduces the differences between traditional methods and the variable population approach. Subsequently, three metrics are designed as criteria to implement a population size adjustment mechanism. This mechanism includes an individual addition mechanism and an individual deletion mechanism. Experimental results show that the variable population method can significantly reduce the number of fullwave simulations required for antenna optimization. © 2024 IEEE.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Improving power system performance by integrating TCSC and SVC devices in the presence of stochastic renewable energy generators

引用

WIND ENGINEERING 2025年

作者： Benhabsa, Imadeddine Kouadri, Ramzi Slimani, Linda Mouassa, Souhil Univ Set 1 Elect Engn Dept Bez Campus Setif 19137 Algeria Univ Bouira Elect Engn Dept Bouira Algeria

This paper investigates the application of advanced metaheuristic algorithms Blood-Sucking Leech Optimizer (BSLO), Bonobo Optimizer (BO), and Electric Eel Foraging optimization (EEFO) to solve the optimal power flow (OPF) problem with stochastic renewable energy generators (REGs), specifically photovoltaic power generators (PVGs) and wind power generators (WGs). Two scenarios are examined: Scenario 1 evaluates the proposed algorithms performance without Flexible AC Transmission Systems (FACTS), focusing on minimizing Total Generation Cost (TGC), Active Power Loss (APL), and a combined objective of TGC and Emissions (TGCE). The TGC including both thermal and REG costs, in which the cost related to stochastic PV and wind power generation encompasses direct, reserve, and penalty costs due to overestimation and underestimation of available PV and wind power. Scenario 2 introduces the Thyristor-Controlled Series Capacitor (TCSC) and the Static Var Compensator (SVC) to evaluate their impact on three objective functions. The performance of the algorithms is evaluated on a modified IEEE 30-bus system. The results show that the BSLO algorithm consistently achieves the best TGC, APL, and TGCE values at 781.1209 $/h, 1.9960 MW, and 810.7376 $/h, respectively. These outcomes highlight its effectiveness and competitive performance in the first scenario. The integration of FACTS devices in the second scenario results in a 6.73% reduction in APL with the insertion of TCSC, a 1.86% reduction with the insertion of SVC, and a 6.10% reduction with the insertion of both TCSC and SVC, compared to the APL value in the case without FACTS devices (1.9960 MW). The study comprehensively analyzes how different optimization techniques and FACTS devices enhance power system performance with stochastic REG integration.

关键词： Optimal power flow stochastic PV and wind power generation thyristor-controlled series capacitor the static var compensator blood-sucking leech optimizer optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Reliability-Aware Network Slicing based on Multi-Objective optimization 13

Reliability-Aware Network Slicing based on Multi-Objective O...

引用

13th IEEE Congress on Evolutionary Computation, CEC 2024

作者： Xia, Qiqi Wang, Jin Huang, Chengqiang Yao, Xin Southern University of Science and Technology Shenzhen China Southern University of Science and Technology Guangdong Key Laboratory of Brain-inspired Intelligent Computation Department of Computer Science and Engineering Shenzhen China Huawei Technologies Co. Ltd. Reliability Technology Lab Shenzhen China Lingnan University Department of Computing and Decision Sciences Hong Kong

ISBN: (纸本)9798350308365

Network slicing is an essential technology in 5G and the forthcoming 6G networks. It aims to embed multiple virtual networks, i.e., network slices, on top of a shared substrate network to meet diverse service requirements. While a considerable body of existing research strives to maximize overall profits by meeting the resource demands of the network slices, optimizing their reliability is frequently overlooked. In this paper, we formalize the network slicing problem as a multi-objective optimization problem that aims to maximize total profits and reliability of network slices. To tackle this problem, we propose a new multi-objective optimization approach that improves over the state-of-the-art algorithm, which can achieve good approximate Pareto front results balancing total profits and reliability of network slices. The performance of our proposed method is evaluated on both artificial and real-world network topologies. Experimental results demonstrate the superior performance of our proposed method compared to the baseline algorithm, outperforming the latter in 92% of instances in terms of the Hypervolume (HV) metric. © 2024 IEEE.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

A PRELIMINARY STUDY ON ACCELERATING SIMULATION optimization WITH GPU IMPLEMENTATION

A PRELIMINARY STUDY ON ACCELERATING SIMULATION OPTIMIZATION ...

引用

2024 Winter Simulation Conference

作者： He, Jinghai Liu, Haoyu Wu, Yuhang Zheng, Zeyu Zhu, Tingyu Univ Calif Berkeley Dept Ind Eng & Operat Res Berkeley CA 94720 USA

ISBN: (纸本)9798331534202;9798331534219

We provide a preliminary study on utilizing GPU (Graphics Processing Unit) to accelerate computation for three simulation optimization tasks with either first-order or second-order algorithms. Compared to the implementation using only CPU (Central Processing Unit), the GPU implementation benefits from computational advantages of parallel processing for large-scale matrices and vectors operations. Numerical experiments demonstrate such computational advantages of utilizing GPU implementation in simulation optimization problems, and show that such advantage comparatively further increase as the problem scale increases.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Di-AEFA: Diversified Artificial Electric Field Algorithm for Global Numerical optimization 8

Di-AEFA: Diversified Artificial Electric Field Algorithm for...

引用

8th International Symposium on Innovative Approaches in Smart Technologies, ISAS 2024

作者： Feda, Afi Kekeli Adegboye, Oluwatayomi Rereloluwa Tibetan, Abosede Omowumi Ojekemi, Opeoluwa Seun Oz, Tolga Ulker, Ezgi Deniz European University of Lefke Advanced Research Centre Mersin-10 Turkey Management Information System University of Mediterranean Karpasia Mersin-10 Turkey Occupational Health and Safety University of Mediterranean Karpasia Mersin-10 Turkey Engineering Management University of Mediterranean Karpasia Mersin-10 Turkey University of Mediterranean Karpasia Aviation Management Department Mersin-10 Turkey European University of Lefke Mersin-10 Turkey

ISBN: (纸本)9798331540104

The Artificial Electric Field Algorithm (AEFA) is categorized as one of the strong algorithms in the area of metaheuristics. It is influenced by Coulomb's law of the electrostatic force. Although it is known as an acceptable algorithm, AEFA is susceptible to early convergence because of its inherent properties, such as the strong attraction force that exist between particles. This paper proposes a diversified AEFA algorithm (Di-AEFA) which is a variation of the original AEFA to form a solid balance between exploration and exploitation. Using a set of 17 strong static benchmark functions, Di-AEFA was assessed and contrasted against a variety of state-of-art algorithms, such as Particle Swarm optimization (PSO), Cuckoo Search (CS), Firefly Algorithm (FA), Salp Swarm Algorithm (SSA) and Differential Evolution (DE). Experimental and statistical results demonstrate that Di-AEFA produces quite acceptable results and performs better than the algorithms tested. Additionally, the performance of Di-AEFA is validated using five dynamic benchmark functions, and it was compared with six powerful algorithms. Overall, the results reveal that Di AEFA performs better than the original AEFA and the counterpart algorithms © 2024 IEEE.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Water resource vulnerability assessment in Hubei Province: a case study

引用

Energy Informatics 2024年第1期7卷 119页

作者： Li, Qiong Zhou, Jian Zhang, Zhinan School of Mathematics and Physics Hubei Polytechnic University Huangshi435000 China Huangshi435000 China Wuhan Raycus Fiber Laser Technologies Co. LTD Wuhan430074 China

In view of the different views of academia on the weight allocation of vulnerability assessment indicators, this study creatively proposed a data-based objective evaluation framework of water resource vulnerability, and applied it to the evaluation of water resource vulnerability in Hubei Province. According to the conceptual model of DPSIR proposed by the United Nations, five vulnerability factors are proposed: driving force, pressure, state, influence and response. In this study, 15 indicators were selected and the projection tracing model was used to identify vulnerability. Aiming at the complex problem of optimization calculation of projection index function in the projection tracing model, the accelerated genetic algorithm is used to speed up the optimization speed, solves the optimization problem in the process of projection tracing, and determines the objective weight of all indicators. Example calculation shows that the model can deal with complex multi-index optimization problems, and is an effective way to solve the comprehensive evaluation of complex vulnerability, and the weighting method is important for the evaluation of water resources vulnerability. The results of this paper show that the combination of projection tracing method and machine learning algorithm can improve the efficiency, objectivity and accuracy of high-dimensional data analysis, and can provide scientific basis for policy makers. © The Author(s) 2024.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

A Transparent and Flexible Polarization Conversion Electromagnetic Scatterer

引用

IEEE ANTENNAS AND WIRELESS PROPAGATION LETTERS 2024年第12期23卷 4912-4916页

作者： Zhen, Qiutong Mao, Zhoutao Chen, Hongzhi Chen, Jing Chernogor, Leonid F. Jin, Zhejun Sun, Zhongsen Zheng, Yu Zhou, Peiheng Liu, Tian Qingdao Univ Coll Elect Informat Qingdao 266071 Peoples R China Univ Elect Sci & Technol China Natl Engn Res Ctr Electromagnet Radiat Control Mat Key Lab Multispectral Absorbing Mat & Struct Minist Educ Chengdu 610054 Peoples R China Jilin Univ Coll Elect Sci & Engn Int Ctr Future Sci State Key Lab Integrated Optoelect Changchun 130012 Peoples R China

This letter presents a novel design for a transparent and flexible electromagnetic scatterer aimed at reducing radar cross section (RCS). Made from low-resistance indium tin oxide (ITO) thin film, PET, and PVC plastic, it achieves up to 81.3% transparency. The innovative unit design overcomes ITO's conductivity limitations and broadens bandwidth. Using the Pancharatnam-Berry phase principle, rotating the top-layer conductive structure creates encoded supercells. The letter explores RCS optimization using multiple algorithms, analyzes RCS changes in metasurface scatterer (MSS) with ITO patches in the conformal state, and compares the effects of ITO and copper on RCS performance. Multialgorithm optimization results in over 10 dB RCS reduction from 8.5 GHz to 19.8 GHz, even under 2 pi curvature or 45 degrees oblique incidence. With a compact size (0.135 lambda(0)), transparency(81.3%), flexibility(central angle up to 2 pi), and wideband performance (80%), the scatterer is ideal for applications like multispectral scattering, antenna radomes, and transparent windows.

关键词： Indium tin oxide Scattering Metasurfaces Wideband Reflection optimization Encoding Simulation Reflection coefficient Bending optimization algorithms Pancharatnam-Berry (PB) phase polarization conversion radar cross section (RCS) transparent and flexible scatterer

来源：评论

学校读者我要写书评

暂无评论

A hybrid learning approach for simulating settlement of shallow foundation

引用

MULTISCALE AND MULTIDISCIPLINARY MODELING EXPERIMENTS AND DESIGN 2025年第1期8卷 1-15页

作者： Li, Jiaman Wu, Jundong Hu, Wei Shaanxi Coll Commun Technol Xian 710000 Shaanxi Peoples R China China Airport Planning & Design Inst Co Ltd Northwest Branch Xian 710000 Shaanxi Peoples R China Shaanxi Prov Transport Planning Design & Res Inst Xian 710065 Shaanxi Peoples R China

Settlement modeling is essential because of the cohesive soil texture's intricacy. This study seeks to identify settlement (Sm) of shallow foundations using newly discovered ML approaches named the hybridized random forests analysis (RF) with grasshoppers optimization algorithm (GOA), the bat-inspired approach (BAT), beluga whale optimization (BWH). RF serves as the primary predictive model due to its robust ensemble learning approach. optimization algorithms enhance its predictive accuracy. Hybrid models (GOARF, BATRF, BWHRF) leverage the strengths of both RF and optimization algorithms to provide superior settlement predictions. By combining RF with these optimization techniques, the study aims to achieve highly accurate predictions of settlement for shallow foundations, demonstrating the effectiveness of these advanced machine-learning approaches. All the BATRF, GOARF, and BWHRF models accurately emulated the Sm, with R2 values of at least 0.985 for the training and 0.978 for the testing collection, respectively. Comparing the BWHRF to other models and literature, it is believed to be the appropriate system with the highest categorization. R2, RMSE, and MAE values during learning are 0.9913, 2.341, and 1.239 mm, respectively, which are superior by 0.9025, 8.09, and 4.92 mm over ANFIS-PSO. The values of the A_(15-index) index depict that the BWHRF can be outperformed by other models. Finally, after examining the validity and considering the assumptions, it is clear that the RF paired with BWH can perform more effectively than the BAT and GOA, and even ANFIS-PSO (literature), in the Sm simulation.

关键词： Settlement Shallow foundation Estimation Random forests optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Towards self-organizing railway traffic management: concept and framework

引用

JOURNAL OF RAIL TRANSPORT PLANNING & MANAGEMENT 2024年 29卷

作者： D'Amato, Leo Naldini, Federico Tibaldo, Valentina Trianni, Vito Pellegrini, Paola Politecn Torino Dipartimento Automat & Informat Corso Duca degli Abruzzi 24 Turin Italy CNR Inst Cognit Sci & Technol Via San Martino della Battaglia 44 I-00185 Rome Italy Univ Gustave Eiffel COSYS ESTAS F-59650 Villeneuve Dascq France

Railway traffic management requires a timely and accurate redefinition of routes and schedules in response to detected perturbations of the original timetable. To date, most of the (automated) solutions to this problem require a central authority to make decisions for all the trains in a given control area. An appealing alternative is to consider trains as intelligent agents able to self -organize and determine the best traffic management strategy. This could lead to more scalable and resilient approaches, that can also take into account the real-time mobility demand. In this paper, we formalize the concept of railway traffic self -organization and we present an original design that enables its real -world deployment. We detail the principles at the basis of the sub -processes brought forth by the trains in a decentralized way, explaining their sequence and interaction. Moreover, we propose a preliminary proof of concept based on a realistic setting representing traffic in a French control area. The results allow conjecturing that self -organizing railway traffic management may be a viable option, and foster further research in this direction.

关键词： Railway traffic management optimization algorithms Self-organization Consensus

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：