检索结果-内蒙古大学图书馆

41st International Conference on Machine Learning (ICML)

作者： Amortila, Philip Foster, Dylan J. Krishnamurthy, Akshay Univ Illinois Urbana IL 61801 USA Microsoft Res Redmond WA USA

Exploration is a major challenge in reinforcement learning, especially for high-dimensional domains that require function approximation. We propose exploration objectives-policy optimization objectives that enable downstream maximization of any reward function-as a conceptual framework to systematize the study of exploration. We introduce a new objective, L-1-Coverage, which generalizes previous exploration schemes and supports three fundamental desiderata: 1. Intrinsic complexity control. L-1-Coverage is associated with a structural parameter, L-1-Coverability, which reflects the intrinsic statistical difficulty of the underlying MDP, subsuming Block and Low-Rank MDPs. 2. Efficient planning. For a known MDP, optimizing L-1-Coverage efficiently reduces to standard policy optimization, allowing flexible integration with off-the-shelf methods such as policy gradient and Q-learning approaches. 3. Efficient exploration. L-1-Coverage enables the first computationally efficient model-based and model-free algorithms for online (reward-free or reward-driven) reinforcement learning in MDPs with low coverability. Empirically, we find that L-1-Coverage effectively drives off-the-shelf policy optimization algorithms to explore the state space.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

COMPARING GRADIENT-FREE AND GRADIENT-BASED MULTI-OBJECTIVE optimization METHODOLOGIES ON THE VKI-LS89 TURBINE VANE TEST CASE

COMPARING GRADIENT-FREE AND GRADIENT-BASED MULTI-OBJECTIVE O...

引用

67th ASME Turbomachinery Technical Conference and Exposition (Turbo Expo) on Road Mapping the Future of Propulsion and Power

作者： Hottois, Romain Chatel, Arnaud Coussement, Gregory Verstraete, Tom De Bruyn, Tom von Karman Inst Turbomachinery & Prop Dept Rhode St Genese Belgium Univ Mons Fac Polytech Mons Fluides Machines Dept Mons Belgium UGent Fac Engn & Architecture Dept Electromech Syst & Met Engn Ghent Belgium

ISBN: (纸本)9780791886120

The present paper addresses the multi-objective aerodynamic shape optimization of the two-dimensional LS-89 turbine cascade. The objective is to minimize the entropy generation at subsonic and transonic flow conditions while maintaining the same flow turning. Nineteen design variables are used to parametrize the geometry. The optimization problem is used to compare two major classes of optimization algorithms and at the same time deduce if this problem has multiple local solutions or one global optimum. A first optimization strategy uses a gradient-based Sequential Quadratic Programming algorithm. This SQP algorithm allows to directly handle the non-linear constraints during the optimization process. An adjoint solver is used for computing the sensitivities of the flow quantities with respect to the design variables, such that the additional gradient computational cost is nearly independent of the number of design variables. In addition, the same optimization problem is performed with a gradient-free - metamodel assisted - evolutionary algorithm. A comparison of the two Pareto-fronts obtained with both methods shows that the gradient-based approach allows to find the same optimum at a reduced computational cost. Moreover, the results suggest that the considered optimization problem is uni-modal. In other terms, it is characterized by a single optimal solution.

关键词： Turbines Pareto optimization optimization Design algorithms Flow (Dynamics) Entropy Cascades (Fluid dynamics) Evolutionary algorithms Geometry optimization algorithms Quadratic programming Transonic flow Flow turning Shape optimization

来源：评论

学校读者我要写书评

暂无评论

Application and optimization of residual connection neural network in spacecraft thermal design

引用

CASE STUDIES IN THERMAL ENGINEERING 2024年 61卷

作者： Hu, Jiangcheng Guo, Liang Zheng, Weihan Chinese Acad Sci Changchun Inst Optics Fine Mech & Phys Changchun 130033 Jilin Peoples R China Univ Chinese Acad Sci Beijing Peoples R China

In thermal analysis modeling, the finite element method (FEM) is commonly used;however, it incurs high computational costs and complicates the global optimization of thermal parameters. To address these challenges, developing simplified surrogate models is crucial for enhancing analysis efficiency. Yet, constructing such models demands exceptional predictive accuracy, making conventional parameter adjustment methods inadequate for design needs. This paper introduces a novel Residual connection Neural Network model, called Res-NN, designed to approximate the CMOS finite element model. By employing residual connections, the Res-NN model significantly reduces fitting errors between networks, achieving a predictive accuracy of 94.6 %, which is 6.6 % higher than that of comparable Multi-Layer Perceptron (MLP) models. Moreover, Res-NN is over 100 times faster than traditional FEM in prediction speed, effectively circumventing the difficulties associated with parameter adjustments. To optimize the temperature fluctuations of the CMOS model, we utilized the Res-NN model as an iterative object within an optimization algorithm. Through experimental comparisons, we identified the PSO algorithm as the most effective option. The PSO optimization results demonstrated a chip temperature difference of 0.045 degrees C, with a simulation error of only 0.0649 degrees C, meeting design specifications and achieving a 61.7 % reduction in temperature variance compared to traditional thermal design method. This validates the superiority of the entire optimization process.

关键词： Surrogate model Residual connected Neural network optimization algorithms Spacecraft thermal design

来源：评论

学校读者我要写书评

暂无评论

Advancements in Bat algorithm 4th

Advancements in Bat algorithm

引用

4th International Conference on Computational Methods in Science and Technology, ICCMST 2024

作者： Kaur, Arvinder Kumar, Yugal Sidhu, Jagpreet Department of Information Technology Chandigarh Engineering College-CGC Landran Punjab Mohali India Computer Engineering/STME NMIMS Chandigarh Campus India

ISBN: (纸本)9781032911571

The modern times have led to the adoption of distinctive meta-heuristic procedures for solving distinct class of optimization-problems. The meta-heuristics procedures have benefit above conventional algorithms because of strong capabilities of exploration and exploitation. In a similar vein, a Bat algorithm was created based on the echolocation habits of bats to address optimisation difficulties. Several advancements have been made in bat algorithm since then to achieve the state-of art results. The purpose of the study is to investigate bat algorithm for the various advancements done. Henceforth, the study presents the miniature analysis on advancements in bat algorithm. The bat algorithm applications in different areas are also investigated. © 2025 the Author(s).

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

A Graph Clustering Algorithm Based on Gravitational Vector 16th

A Graph Clustering Algorithm Based on Gravitational Vector

引用

16th International Conference on Genetic and Evolutionary Computing, ICGEC 2024

作者： Li, Jitao Yao, Yao Xu, Chugui Tian, Xue Department of Information Science and Technology Taishan University Tai’an271000 China

ISBN: (纸本)9789819615346

The traditional clustering algorithms mostly use simple one-dimensional distance measurement and overall optimization strategy for classification, one-dimensional measurement method can identify similarity matrix, but cannot keep the spatial location relationship between samples, because of the loss of data information between the sample which cause the clustering accuracy problem, this paper puts forward a graph clustering algorithm based on gravitational vector. This method combines the principle of gravity principle, uses the multi-dimensional vector instead of the traditional one-dimensional distance measurement method, and excavates the connection between the samples by establishing the local balance relationship, and finally achieves the purpose of improving the local information extraction and data distribution perception ability of the algorithm. Using the present algorithm in artificial datasets and UCI datasets can show that the present algorithm can cope well with multi-type, unbalanced, high complexity data and get good clustering results. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Maximizing Throughput and Minimizing Latency: Evolutionary algorithms for Optimizing Wireless Cellular Networks 3

Maximizing Throughput and Minimizing Latency: Evolutionary A...

引用

3rd International Conference on Intelligent Data Communication Technologies and Internet of Things, IDCIoT 2025

作者： Pechetti, Jyoshitha Yadav, Sneha Tarunya, G. Rajagopal, Shinu M. Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Bengaluru India

ISBN: (纸本)9798331527549

In the rapidly evolving world of wireless cellular network, optimizing key parameters like data throughput and latency is of critical importance for ensuring high quality communication services. The proposed presents a novel approach to multi-objective optimization to enhance overall network performance using Genetic algorithms (GA), Pareto, and Particle Swarm optimization (PSO). By employing optimization algorithms, a balance is sought in terms of trade-off of other parameters while maximizing data throughput and minimizing latency. Simulation layers indicate the comparative study further solidifies the efficacy of the value of evolutionary algorithms in optimizing wireless networks to satisfy the growing needs of high speed and less latency services in today’s communication landscape. © 2025 IEEE.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Predicting the seismic response of the short structures by considering the Whale optimization Algorithm

引用

ENERGY REPORTS 2021年 7卷 4071-4084页

作者： Liu, Zhe Zhang, Lei Li, Jiajing Mamluki, Mina Guangdong Univ Petrochem Technol Sch Petr Engn Maoming 525000 Guangdong Peoples R China Chongqing Univ Sci & Technol Sch Petr Engn Chongqing 401331 Peoples R China Univ Mohaghegh Ardabili Ardebil Iran

This paper presents a newly improved version of the Whale optimization Algorithm based on quantum theory by optimizing the Artificial Neural Network for the ground response approximation problems in short buildings. Based on the literature, Artificial Neural Network does not provide a proper response of ground surface prediction in seismic load applications. Hence, the main idea of this study is to assess the utilization of the whale optimization algorithm for calculating the columns' horizontal deflection in the short structure under notable considered seismic loading. The input database utilized in this paper is the Chi-Chi earthquake that is occurred in 1999 in Taiwan. To provide the train and test databases for the artificial neural network algorithm and whale optimization algorithm, Finite Element models are used. The inputs contain the Chi-Chi earthquake's dynamic time, soil elastic modulus, dilation angle, Poisson's ratio, unit weight, friction angle, bending stiffness and axial stiffness. Also the columns' horizontal deflection in their highest level is considered as the output. The results showed the higher reliability of the proposed WOA in calculating the ground response and the column's horizontal deflection in short buildings subjected to the earthquake loading. (C) 2021 The Authors. Published by Elsevier Ltd.

关键词： Short structures Earthquake ANN WOA-ANN PSO-ANN optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Multi-Revolution Optimal Rendezvous with Pertubation Near Elliptical Reference Orbit

Multi-Revolution Optimal Rendezvous with Pertubation Near El...

引用

2024 IEEE International Conference on Control Science and Systems Engineering, ICCSSE 2024

作者： Chen, Weiyue Qiu, Jinjuan China Academy of Space Technology Beijing China Beijing Institute of Control Engineering China Academy of Space Technology Beijing China

ISBN: (纸本)9798331517199

For multi-revolution rendezvous near elliptical reference orbit, the singularities of the semi-revolution and complete-revolution rendezvous are eliminated. The semi-revolution and complete-revolution rendezvous schemes are proposed, and an implementable rough-refinement two-step optimization algorithm is proposed to determine the optimal transfer time and two impulses. Based on the optimization process, a differential gradient iteration algorithm is presented to improve the rendezvous accuracy with the orbit perturbation considered. The multi-revolution two-impulse rendezvous optimization algorithm and iteration guidance algorithm can be easily extended to multiple-impulse rendezvous. Simulation examples of rendezvous near elliptical reference orbits with different eccentricities are performed. The results show that the optimization algorithm is accurate and effective, and the iteration algorithm converges rapidly. It is concluded that the optimization algorithm and the iteration guidance algorithm are suitable for perturbed multi-revolution rendezvous near circle reference orbit and arbitrary elliptical reference orbit. Several iterations can make the terminal rendezvous error less than one meter. © 2024 IEEE.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Speeding up NAS with Adaptive Subset Selection 3

Speeding up NAS with Adaptive Subset Selection

引用

International Conference on Automated Machine Learning (AutoML)

作者： Prasad, Vishak C. White, Colin Nayak, Sibasis Jain, Paarth Shameem, Aziz Garg, Prateek Ramakrishnan, Ganesh Indian Inst Technol Bombay Maharashtra India Abacus AI San Francisco CA USA

A majority of recent developments in neural architecture search (NAS) have been aimed at decreasing the computational cost of various techniques without affecting their final performance. Towards this goal, several low-fidelity and performance prediction methods have been considered, including those that train only on subsets of the training data. In this work, we present an adaptive subset selection approach to NAS and present it as complementary to state-of-the-art NAS approaches. We uncover a natural connection between one-shot NAS algorithms and adaptive subset selection and devise an algorithm that makes use of state-of-the-art techniques from both areas. We use these techniques to substantially reduce the runtime of DARTS-PT (a leading one-shot NAS algorithm) with minimal sacrifice to accuracy and to give reasonable runtime reductions for BOHB and DEHB (leading multi-fidelity optimization algorithms) as well. In experiments, we find architectures on CIFAR-10 that give 5% increase in performance over DARTS-PT while reducing the time required by more than a factor of 8. Our results are consistent across multiple datasets, and towards full reproducibility, we release all our code at https://***/pcvishak/SubsetSelection_NAS.

关键词： optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Assessing CNN’s Performance with Multiple optimization Functions for Credit Card Fraud Detection

引用

Procedia Computer Science 2024年 235卷 2035-2042页

作者： Chandana Gouri Tekkali Karthika Natarajan Research Scholar School of Computer Science and Engineering VIT-AP University Amaravati Andhra Pradesh India Assistant Professor School of Computer Science and Engineering VIT-AP University Amaravati Andhra Pradesh India

In recent days, credit card fraud has emerged as a significant challenge for researchers in the field of detection and prevention. Tackling this challenge holds substantial benefits for both public and private organizations, as it directly impacts economic statistics. Our proposed model introduces crucial advancements to address this issue in real-time scenarios. To achieve this, we harnessed the power of Deep Convolutional Neural Networks (DeepConvNet) in conjunction with various optimization techniques. optimization algorithms encompass a group of mathematical and computational techniques employed to discover the most suitable solution or set of solutions for a given problem. The primary goal of these algorithms is to optimize, either by maximizing or minimizing, an objective function while ensuring compliance with specific constraints. In this research work, we provide a comparison of highly effective and validated optimization techniques: Stochastic Gradient Descent (Sgd), Adaptive Gradient (Adagrad), Adaptive Moment Estimation (Adam), and Root Mean Squared Propagation (Rmsprop). These optimization algorithms are applied to the Deep Convolutional Neural Network (DeepConvNet) in the subject of our specific problem statement, which involves credit card fraud detection (CCFD). After careful consideration of the problem’s nature, objective function characteristics, and computational aspects, we fnd that all four algorithms are suitable for our CCFD task. However, based on experimental results, it is evident that Rmsprop outperforms others, leading to a remarkable 99.93% in accuracy.

关键词： optimization algorithms Objective functions Convolutional Neural Network Credit Card Fraud Detection

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：