检索结果-内蒙古大学图书馆

Evolving filter criteria for randomly initialized network pruning in image classification

NEUROCOMPUTING 2024年 594卷

作者： Chen, Xiangru Liu, Chenjing Hu, Peng Lin, Jie Gong, Yunhong Chen, Yingke Peng, Dezhong Geng, Xue Sichuan Univ Coll Comp Sci Chengdu 610065 Peoples R China Agcy Sci Technol & Res Inst Infocomm Res Singapore 138632 Singapore Sichuan Newstrong UHD Video Technol Co Ltd Chengdu 610095 Peoples R China Northumbria Univ Dept Comp & Informat Sci Newcastle Upon Tyne NE1 8ST England

Network pruning aims to enhance the performance of deep neural networks by eliminating redundant components from the model. However, existing pruning methods typically require a well -trained model and employ fixed, single pruning criteria throughout the pruning cycles. To address these limitations, we propose a novel method called evolutionary Filter Criteria (EvoFC). This method enables the automated search for the network pruning ratio and criterion during a population -based heuristic search process. We introduce a unique encoding space that represents the chosen pruning criterion and ratio for each layer, facilitating the acquisition of optimal architecture configurations for candidate networks during iterations. Additionally, we devise a novel weight inheritance mechanism to mitigate the computational burden associated with the population -based nature of the method, resulting in a significant reduction in overall training time. We validate our method by applying it to randomly initialized networks and conducting empirical experiments on CIFAR-10/100, ILSVRC2012 and Places365 datasets. The results demonstrate that our method effectively reduces the number of FLOPs while striking a fine balance between accuracy and computational efficiency. This underscores the practical value of our method in optimizing performance while efficiently utilizing computational resources, particularly when pruning networks starting from random initialization.

关键词： Genetic algorithm Evolving deep learning Network pruning multi-objective evolutionary algorithm

来源：评论

学校读者我要写书评

暂无评论

multi-objective boxing match algorithm for multi-objective optimization problems

引用

EXPERT SYSTEMS WITH APPLICATIONS 2024年 239卷

作者： Tavakkoli-Moghaddam, Reza Akbari, Amir Hosein Tanhaeean, Mehrab Moghdani, Reza Gholian-Jouybari, Fatemeh Hajiaghaei-Keshteli, Mostafa Univ Tehran Sch Ind Engn Coll Engn Tehran Iran Iran Univ Sci & Technol Sch Ind Engn Tehran Iran Univ Huddersfield Sch Business & Educ Dept Accountancy Finance & Econ Queensgate Rd Huddersfield HD1 3DH W Yorkshire England Tecnol Monterrey Sch Engn & Sci Puebla Mexico

In the last two decades, due to having fast computation after inventing computers and also considering realworld optimization problems, research on developing new algorithms for problem having more than one objective have been one of the appealing and attractive topics both for academia and industrial practitioners. By this motivation, we introduce a multi-objective Boxing Match algorithm (MOBMA) in this paper. The proposed algorithm studies the multi-objective version of the Boxing Match algorithm (BMA) by incorporating a unique search strategy and new solutions-producing mechanism, enhancing the algorithm's capability for exploration and exploitation phases. Besides, its performance is analyzed with famous and capable multi-objective metaheuristics. We consider ten multi-objective benchmarks and three classical engineering problems. Statistical analyses are also conducted on the benchmark test functions from three engineering design problems. This study shows the superior performance of the proposed algorithm, considering both quantitative and qualitative analyses.

关键词： Meta-heuristics multi-objective optimization Boxing match algorithm multi-objective evolutionary algorithm

来源：评论

学校读者我要写书评

暂无评论

Solving multi-objective weapon-target assignment considering reliability by improved MOEA/D-AM2M

引用

NEUROCOMPUTING 2024年 563卷

作者： Yi, Xiaojian Yu, Huiyang Xu, Tao Beijing Inst Technol Sch Mech Engn Beijing 100081 Peoples R China Beijing Inst Technol Yangtze Delta Reg Acad Jiaxing 314019 Zhejiang Peoples R China Beijing Inst Technol Tangshan Res Inst Tangshan 063099 Hebei Peoples R China

The weapon-target assignment problem is a challenging optimization issue, but reliability is seldom considered in the majority of existing literature. To address the high-reliability weapon-target assignment problem, this paper integrates weapon reliability and mission reliability into a multi-objective optimization model (MOD) and presents an improved algorithm termed MOEA/D-iAM2M to the problem. This algorithm effectively combines the strengths of adaptive search space decomposition-based MOEA (MOEA/D-AM2M) and two-stage hybrid learning-based MOEA (HLMEA), resulting in a faster convergence rate and a more extensive distribution of the Pareto solution set. Furthermore, a reference point is incorporated into MOEA/D-iAM2M to facilitate the adaptive weight adjustment. Numerical experiments are carried out to confirm the effectiveness of the proposed MOEA/D-iAM2M. This research is significant in the field of optimization and has practical value in the defense industry.

关键词： multi-objective optimization Weapon-target assignment Reliability multi-objective evolutionary algorithm

来源：评论

学校读者我要写书评

暂无评论

Modeling the Optimal Transition of an Urban Neighborhood towards an Energy Community and a Positive Energy District

引用

ENERGIES 2024年第16期17卷 4047页

作者： Viesi, Diego Borelli, Gregorio Ricciuti, Silvia Pernigotto, Giovanni Mahbub, Md Shahriar Fdn Bruno Kessler FBK Ctr Sustainable Energy Via Sommar 18 I-38123 Trento Italy Free Univ Bozen Bolzano UNIBZ Fac Engn Piazza Univ 5 I-39100 Bolzano Italy Ahsanullah Univ Sci & Technol AUST Dept Comp Sci & Engn Love Rd 141&142Tejgaon Ind Area Dhaka 1208 Bangladesh

Building renovation is a key initiative to promote energy efficiency, the integration of renewable energy sources (RESs), and a reduction in CO2 emissions. Supporting these goals, emerging research is dedicated to energy communities and positive energy districts. In this work, an urban neighborhood of six buildings in Trento (Italy) is considered. Firstly, the six buildings are modeled with the Urban Modeling Interface tool to evaluate the energy performances in 2024 and 2050, also accounting for the different climatic conditions for these two time periods. Energy demands for space heating, domestic hot water, space cooling, electricity, and transport are computed. Then, EnergyPLAN coupled with a multi-objective evolutionary algorithm is used to investigate 12 different energy decarbonization scenarios in 2024 and 2050 based on different boundaries for RESs, energy storage, hydrogen, energy system integration, and energy community incentives. Two conflicting objectives are considered: cost and CO2 emission reductions. The results show, on the one hand, the key role of sector coupling technologies such as heat pumps and electric vehicles in exploiting local renewables and, on the other hand, the higher costs in introducing both electricity storage to approach complete decarbonization and hydrogen as an alternative strategy in the electricity, thermal, and transport sectors. As an example of the quantitative valuable finding of this work, in scenario S1 "all sectors and EC incentive" for the year 2024, a large reduction of 55% of CO2 emissions with a modest increase of 11% of the total annual cost is identified along the Pareto front.

关键词： urban modeling interface EnergyPLAN multi-objective evolutionary algorithm energy community positive energy district

来源：评论

学校读者我要写书评

暂无评论

A framework for constrained large-scale multi-objective white-box problems based on two-scale optimization through decision transfer

引用

INFORMATION SCIENCES 2024年 665卷

作者： Wang, Qingzhu Li, Tianyang Meng, Fanqi Li, Bin Northeast Elect Power Univ Sch Comp Sci Jilin 132012 Peoples R China Guangdong Atv Acad Performing Arts Sch Informat Engn Dongguan 523710 Peoples R China

Most existing constrained multi-objective evolutionary algorithms (CMOEAs) are not so efficient when handling constrained large-scale multi-objective problems (CLSMOPs). To overcome whitebox CLSMOPs with definitive objective functions, a two-scale optimization framework based on decision transfer, which integrates dimensionality reduction of large-scale decision variables and constraint handling technology, is proposed. The Lagrange multiplier is first used to construct the two-scale optimization model, which bridges original large-scale decision space of variables and small-scale (2-scale) decision space of objective-constraint parameter. The decision transfer algorithm is then designed to switch between large-scale original decision space and small-scale parametric decision space, while achieving the maximum dimensionality reduction. Finally, the two-scale evolution strategy is proposed for the alternative optimizations in the two decision spaces, which emphasize objectives and constraints, respectively. In summary, the optimization in the large-scale space pushes the population to unconstrained Pareto front (PF), the optimization in the small-scale space helps the population cross the infeasible areas to approach constrained PF, and the offspring generation by Lagrange multiplier is beneficial to both objectives and constraints. Eight representative and state-of-the-art CMOEAs have been embedded into the CLDTEA framework to demonstrate its effectiveness through comparative experiments on CLSMOPs with equality and inequality constraints and 1000 decision variables. Experimental results show that CLDTEA can significantly improve the performance of these basic CMOEAs.

关键词： Constrained large-scale multi-objective problem multi-objective evolutionary algorithm Decision transfer Lagrange multiplier

来源：评论

学校读者我要写书评

暂无评论

Kinetic facades: An evolutionary-based performance evaluation framework

引用

JOURNAL OF BUILDING ENGINEERING 2022年 53卷

作者： Sadegh, Salman Oukati Gasparri, Eugenia Brambilla, Arianna Globa, Anastasia Politecn Milan Dept Architecture Built Environm & Construction En Milan Italy Univ Sydney Sch Architecture Design & Planning Sydney Australia

multi-functional kinetic facades have great potential to adapt to the environment, reduce building energy consumption, regulate shading and natural ventilation, and improve human thermal and visual comfort. However, due to their complex nature, kinetic facades systems are often associated with highly onerous time-consuming design and fabrication processes, which significantly limit their diffusion and market uptake. This research aims to propose an integrated, systematic framework that supports kinetic facades' performance design and decision making since projects' early design stages. The proposed method makes use of multi-objective evolutionary algorithms (MOEA) which have proven successful in providing optimum solutions for complex problems. The efficacy of the framework is then validated based on a case study analysis to improve daylighting performance in an office building context. The results indicate that the integration of the proposed framework can provide a variety of data to support performance evaluation and to identify the most efficient kinetic facades solutions;moreover, this method provided a variety of optimal solutions using Pareto front and the Ranking Method;each being advantageous in a specific set of criteria as evidenced by improved useful daylight illuminance by up to 30%, daylight autonomy by up to 20% while keeping the under-lit and over-lit values as low as under 10%. Finally, future research can focus on the integration of soft computing methods to create a smart and intelligent performance evaluation framework for kinetic facades.

关键词： Adaptive building envelope Responsive building fa?ade multi-objective evolutionary algorithm Evaluation framework Building performance

来源：评论

学校读者我要写书评

暂无评论

Hybrid loading situation vehicle routing problem in the context of agricultural harvesting: A reconstructed MOEA/D with parallel populations

引用

SWARM AND evolutionary COMPUTATION 2024年 91卷

作者： Guo, Xiang Miao, Zhong-Hua Pan, Quan-Ke He, Xuan Shanghai Univ Sch Mechatron Engn & Automat Shanghai 200444 Peoples R China Liaocheng Univ Sch Comp Sci Liaocheng 252000 Peoples R China Shanghai Maritime Univ Sch Logist Engn Shanghai 201306 Peoples R China

With the increasing level of agricultural automation, the combination of agriculture and intelligent vehicle technology is propelling the development of smart agriculture. Although this technology has already been applied widely for various agricultural production tasks, inefficient vehicle scheduling still hasn't been resolved satisfactorily. Oriented towards agricultural harvesting scenarios, a hybrid loading situation vehicle routing problem (HLSVRP) model is proposed to minimize total energy consumption and maximum completion time. A reconstructed multi-objective evolutionary algorithm based on decomposition (R-MOEA/D) is developed to solve the problem. Eight solution representations tailored specifically to the problem are introduced by R-MOEA/D, allowing an extensive exploration of the solution space. A modified Clarke & Wright (MCW) heuristic is designed to generate a high-quality initial population. A novel problem-specific parallel population updating mechanism based on the four crossover and two mutation combinations is also provided to improve the exploration ability. A collaborative search strategy is employed to facilitate cooperation among parallel populations. Finally, a series of comparative experiments conducted on various task scales and vehicle scales verify the effectiveness of the proposed algorithmic components and the exceptional performance for solving HLSVRP.

关键词： Agricultural harvesting Hybrid loading situations multi-objective evolutionary algorithm Problem-specific parallel population Collaborative search

来源：评论

学校读者我要写书评

暂无评论

NSGA-II/EDA Hybrid evolutionary algorithm for Solving multi-objective Economic/Emission Dispatch Problem

引用

ELECTRIC POWER COMPONENTS AND SYSTEMS 2018年第10期46卷 1158-1170页

作者： Alawode, Kehinde Olukunmi Adegboyega, Gabriel Adebayo Muhideen, Jubril Abimbola Osun State Univ Dept Elect & Elect Engn Osogbo Nigeria Obafemi Awolowo Univ Dept Elect & Elect Engn Ife Nigeria

In this study, a hybrid algorithm which combines the NSGA-II with a modified form of the marginal histogram model Estimation of Distribution algorithm (EDA), herein called the NSGA-II/EDA is proposed for solving the multi-objective economic/emission power dispatch problem. The goal is to improve the convergence while preserving the diversity properties of the obtained solution set. The effect of variable interaction on the marginal histogram EDA model is reduced by performing multi-scale Principal Component Analysis on the population of solutions. Also, the concepts of non-domination and elitism have been introduced into the marginal histogram model in order for it to handle multiple objectives. Several optimization runs were carried out on the standard multi-objective test problems, including the IEEE 30- and the 118-bus test systems. Standard metrics are used to compare the performance of the developed hybrid approach with that of other multi-objective evolutionary algorithms. The effectiveness of the proposed approach in improved convergence, with good diversity is demonstrated.

关键词： multi-objective evolutionary algorithm hybrid evolutionary algorithm combined economic-emission dispatch problem

来源：评论

学校读者我要写书评

暂无评论

Numerical simulation and microchannels parameters optimization for thermal management of GaN HEMT devices

引用

INTERNATIONAL JOURNAL OF NUMERICAL METHODS FOR HEAT & FLUID FLOW 2021年第9期31卷 2841-2861页

作者： Wang, Jiahao Xia, Guodong Li, Ran Ma, Dandan Zhou, Wenbin Wang, Jun Beijing Univ Technol Beijing Key Lab Heat Transfer & Energy Convers MOE Key Lab Enhanced Heat Transfer & Energy Conse Beijing Peoples R China

Purpose This study aims to satisfy the thermal management of gallium nitride (GaN) high-electron mobility transistor (HEMT) devices, microchannel-cooling is designed and optimized in this work. Design/methodology/approach A numerical simulation is performed to analyze the thermal and flow characteristics of microchannels in combination with computational fluid dynamics (CFD) and multi-objective evolutionary algorithm (MOEA) is used to optimize the microchannels parameters. The design variables include width and number of microchannels, and the optimization objectives are to minimize total thermal resistance and pressure drop under constant volumetric flow rate. Findings In optimization process, a decrease in pressure drop contributes to increase of thermal resistance leading to high junction temperature and vice versa. And the Pareto-optimal front, which is a trade-off curve between optimization objectives, is obtained by MOEA method. Finally, K-means clustering algorithm is carried out on Pareto-optimal front, and three representative points are proposed to verify the accuracy of the model. Originality/value Each design variable on the effect of two objectives and distribution of temperature is researched. The relationship between minimum thermal resistance and pressure drop is provided which can give some fundamental direction for microchannels design in GaN HEMT devices cooling.

关键词： CFD Optimization multi-objective evolutionary algorithm Microchannel

来源：评论

学校读者我要写书评

暂无评论

Joint sparse neural network compression via multi-application multi-objective optimization

引用

APPLIED INTELLIGENCE 2021年第11期51卷 7837-7854页

作者： Chen, Jinzhuo Xu, Yongnan Sun, Weize Huang, Lei Shenzhen Univ Coll Elect & Informat Engn Guangdong Key Lab Intelligent Informat Proc Shenzhen Guangdong Peoples R China Shenzhen Univ Coll Elect & Informat Engn Shenzhen Key Lab Adv Nav Technol Coll Elect & Inf Shenzhen Guangdong Peoples R China

Over the pass decade, deep neural network (DNN) has been widely applied in various applications. To alleviate the storage and computation requirement of the complicated DNNs, network compression methods are developed. The sparse structure learning methods based on multi-objective optimization have been proven to be valid to balance the sparsity of the network model and network performance. However, when multiple applications are deployed on one single platform simultaneously, these methods become inefficient because each network model for each application needs to be trained and optimized individually. In this article, a multi-objective, multi-application sparse learning model is proposed to optimize multiple targets from a set of applications together. The joint network structure is first proposed. After a pre-training of the network model, a joint multi-objective evolutionary algorithm is derived to solve the optimization problems. Note that an improved initialization method for parent model generation is also developed. Finally, based on the joint loss between the objectives, fine tuning is used to compute the final models with good performance. The proposed method is evaluated under different datasets with a comparison to the state-of-the-art approaches, and experimental results demonstrate that the multi-application optimization model can give much better performance than the single-application optimization ones, especially in the case that different datasets are involved simultaneously.

关键词： Deep neural network Sparse network learning multi-objective evolutionary algorithm multi-objective optimization

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

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

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：