检索结果-内蒙古大学图书馆

32nd ACM International Conference on Information and Knowledge Management (CIKM)

作者： Samizadeh, Mina Tong, Guangmo Univ Delaware Newark DC 19716 USA

ISBN: (纸本)9798400701245

Data-driven approaches have been proven effective in solving combinatorial optimization problems over graphs such as the traveling salesman problems and the vehicle routing problem. The rationale behind such methods is that the input instances may follow distributions with salient patterns that can be leveraged to overcome the worst-case computational hardness. For optimization problems over graphs, the common practice of neural combinatorial solvers consumes the inputs in the form of adjacency matrices. In this paper, we explore a vision-based method that is conceptually novel: can neural models solve graph optimization problems by taking a look at the graph pattern? Our results suggest that the performance of such vision-based methods is not only non-trivial but also comparable to the state-of-the-art matrix-based methods, which opens a new avenue for developing data-driven optimization solvers.

关键词： neural combinatorial optimization Computer Vision

来源：评论

学校读者我要写书评

暂无评论

Rethinking Supervised Learning based neural combinatorial optimization for Routing Problem

引用

ACM Transactions on Evolutionary Learning and optimization 1000年

作者： Shunyu Yao Xi Lin Jiashu Wang Qingfu Zhang Zhenkun Wang School of System Design and Intelligent Manufacturing Southern University of Science and Technology China City University of Hong Kong China Department of Computer Science City University of Hong Kong China School of System Design and Intelligent Manufacturing Southern University of Science and Technology China

neural combinatorial optimization (NCO) is a promising learning-based approach to solving complex combinatorial optimization problems such as the traveling salesman problem (TSP), the vehicle routing problem (VRP), and the orienteering problem (OP). However, how to efficiently train a powerful NCO solver for routing problems remains a crucial challenge. The widely used reinforcement learning method suffers from sparse rewards and low data efficiency, while the supervised learning approach requires a large number of high-quality solutions (i.e., labels) that could be costly to obtain. In this work, we find that simple data augmentation operations can drastically reduce the number of required high-quality solutions for supervised learning. Moreover, simple boosting strategies that leverage the property of multiple optima can significantly improve training efficiency. With only a small set of \(50,000\) labeled instances, supervised learning can achieve a competitive in-distribution performance with the widely-used reinforcement learning counterpart. Furthermore, we also investigate the generalization ability for larger out-of-distribution problems. We believe the findings from this work may lead to a rethinking of the value of data-efficient supervised learning for NCO solver training.

关键词： combinatorial optimization Data Augmentation neural combinatorial optimization Learning to Optimize Supervised Learning

来源：评论

学校读者我要写书评

暂无评论

Multi-Objective neural Evolutionary Algorithm for combinatorial optimization Problems

引用

IEEE TRANSACTIONS ON neural NETWORKS AND LEARNING SYSTEMS 2023年第4期34卷 2133-2143页

作者： Shao, Yinan Lin, Jerry Chun-Wei Srivastava, Gautam Guo, Dongdong Zhang, Hongchun Yi, Hu Jolfaei, Alireza Alibaba Inc Hangzhou 310052 Peoples R China Western Norway Univ Appl Sci N-5063 Bergen Norway Brandon Univ Dept Math & Comp Sci Brandon MB R7A 6A9 Canada China Med Univ Res Ctr Interneural Comp Taichung 40402 Taiwan Asia Univ Coll Informat & Elect Engn Taichung 413 Taiwan Macquarie Univ Dept Comp Sydney NSW 2109 Australia

There has been a recent surge of success in optimizing deep reinforcement learning (DRL) models with neural evolutionary algorithms. This type of method is inspired by biological evolution and uses different genetic operations to evolve neural networks. Previous neural evolutionary algorithms mainly focused on single-objective optimization problems (SOPs). In this article, we present an end-to-end multi-objective neural evolutionary algorithm based on decomposition and dominance (MONEADD) for combinatorial optimization problems. The proposed MONEADD is an end-to-end algorithm that utilizes genetic operations and rewards signals to evolve neural networks for different combinatorial optimization problems without further engineering. To accelerate convergence, a set of nondominated neural networks is maintained based on the notion of dominance and decomposition in each generation. In inference time, the trained model can be directly utilized to solve similar problems efficiently, while the conventional heuristic methods need to learn from scratch for every given test problem. To further enhance the model performance in inference time, three multi-objective search strategies are introduced in this work. Our experimental results clearly show that the proposed MONEADD has a competitive and robust performance on a bi-objective of the classic travel salesman problem (TSP), as well as Knapsack problem up to 200 instances. We also empirically show that the designed MONEADD has good scalability when distributed on multiple graphics processing units (GPUs).

关键词： optimization Evolutionary computation Heuristic algorithms Search problems neural networks Genetics Urban areas Attention mechanism deep reinforcement learning (DRL) multi-objective learning neural combinatorial optimization neural evolutionary algorithm

来源：评论

学校读者我要写书评

暂无评论

A neural combinatorial Approach for School Bus Routing System optimization

A Neural Combinatorial Approach for School Bus Routing Syste...

引用

作者： Qian, Leren Northeastern University

学位级别：Ph.D., Doctor of Philosophy

Maintaining a fleet of buses to transport students to school is a major expense for school districts. An efficient bus routing system can greatly reduce the time and cost associated with transportation, while also ensuring the safety and comfort of students. Optimizing large-scale school bus routing problems (SBRP) is challenging due to the huge number of variables and constraints involved. It is thus treated by decomposing into several subproblems, involving bus stop selection, bus route generation, school bell time adjustment, and bus scheduling. All subproblems of the SBRP are known to be NP-hard. Among all the sub-problems of SBRP, the school bell time adjustment sub-problem has drawn little research attention. Nonetheless, change in the school bell time (start and end time) can be a cost-efficient and practical approach to improve the school bus transportation system. Given a set of bus routes assigned to schools, in this dissertation we develop a multicriteria optimization model that determines the bell times in each school, the assignment of routes to buses, and the schedule of the buses under a single load assumption for both heterogeneous and homogeneous fleet scenarios. The three criteria we consider are (1) minimizing the number of buses, (2) minimizing the total miles of travel distances, and (3) maximizing the preference of all constituents for schools’ bell times. We first introduce an end-to-end neural combinatorial method to efficiently deal with the school bus scheduling problem. Borrowing the idea from the Transformer architecture, an attention model trained using reinforcement learning is applied to the bus scheduling problem by taking the embedded graph information as input, to predict the optimal routing permutation. As a result, a pre-trained model would be available for solving multiple instances simultaneously in parallel. In addition, we design the scheme for generating benchmark datasets for the bus scheduling problem. Then, a 2-stage Tabu

关键词： Deep reinforcement learning neural combinatorial optimization School bell optimization School bus routing Tabu search

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning for Efficient Drone-Assisted Vehicle Routing

引用

APPLIED SCIENCES-BASEL 2025年第4期15卷 2007-2007页

作者： Bogyrbayeva, Aigerim Dauletbayev, Bissenbay Meraliyev, Meraryslan SDU Univ Dept Comp Sci Kaskelen 040900 Kazakhstan

Many exact algorithms, heuristics, and metaheuristics have been proposed to solve the Vehicle Routing Problem with Drones, which involves using a fleet of trucks and drones to fulfil customer orders in last-mile delivery. In this study, the problem is formulated using the Markov Decision Process, and a Reinforcement Learning (RL) based solution is proposed. The proposed RL model is based on an attention-encoder and a recurrent neural network-decoder architecture. This approach enhances coordination by determining which vehicles should visit specific customers and where vehicles can rendezvous, effectively leveraging drones and reducing the overall completion time. The RL model has demonstrated competitive performance compared to benchmark algorithms through extensive experiments.

关键词： reinforcement learning neural combinatorial optimization vehicle routing

来源：评论

学校读者我要写书评

暂无评论

Deep reinforcement learning for machine scheduling: Methodology, the state-of-the-art, and future directions

引用

COMPUTERS & INDUSTRIAL ENGINEERING 2025年 200卷

作者： Khadivi, Maziyar Charter, Todd Yaghoubi, Marjan Jalayer, Masoud Ahang, Maryam Shojaeinasab, Ardeshir Najjaran, Homayoun Univ Victoria Dept Mech Engn Victoria BC V8P 5C2 Canada Univ Victoria Dept Elect & Comp Engn Victoria BC V8P 5C2 Canada Univ Turku Dept Mech & Mat Engn Turku 20014 Finland

Machine scheduling aims to optimally assign jobs to a single or a group of machines while meeting manufacturing rules as well as job specifications. Optimizing the machine schedules leads to significant reduction in operational costs, adherence to customer demand, and rise in production efficiency. Despite its benefits for the industry, machine scheduling remains a challenging combinatorial optimization problem to be solved, inherently due to its Non-deterministic Polynomial-time (NP) hard nature. Deep Reinforcement Learning (DRL) has been regarded as a foundation for "artificial general intelligence"with promising results in tasks such as gaming and robotics. Researchers have also aimed to leverage the application of DRL, attributed to extraction of knowledge from data, across variety of machine scheduling problems since 1995. This paper presents a comprehensive review and comparison of the methodology, application, and the advantages and limitations of different DRL-based approaches. Further, the study categorizes the DRL methods based on the integrated computational components including conventional neural networks, encoder-decoder architectures, graph neural networks and metaheuristic algorithms. Our literature review concludes that the DRL-based approaches surpass the performance of exact solvers, heuristics, and tabular reinforcement learning algorithms in either computation speed, generating near-global optimal solutions, or both. They have been applied to static or dynamic scheduling of different machine environments, which consist of single machine, parallel machine, flow shop, job shop, and open shop, with different job characteristics. Nonetheless, the existing DRLbased schedulers face limitations not only in considering complex operational constraints, and configurable multi-objective optimization but also in dealing with generalization, scalability, intepretability, and robustness. Therefore, addressing these challenges shapes future work in this field.

关键词： Machine scheduling Deep reinforcement learning neural combinatorial optimization Graph neural networks Production scheduling Artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

Improving generalization of neural Vehicle Routing Problem solvers through the lens of model architecture

引用

neural NETWORKS 2025年 187卷 107380页

作者： Xiao, Yubin Wang, Di Wu, Xuan Wu, Yuesong Li, Boyang Du, Wei Wang, Liupu Zhou, You Jilin Univ Coll Comp Sci & Technol Minist Educ Key Lab Symbol Computat & Knowledge Engn Changchun 130012 Peoples R China Nanyang Technol Univ Joint NTU UBC Res Ctr Excellence Act Living Elderl Singapore 639956 Singapore Nanyang Technol Univ WeBank NTU Joint Res Inst Fintech Singapore 639956 Singapore Nanyang Technol Univ Coll Comp & Data Sci Singapore 639956 Singapore

neural models produce promising results when solving Vehicle Routing Problems (VRPs), but may often fall short in generalization. Recent attempts to enhance model generalization often incur unnecessarily large training cost or cannot be directly applied to other models solving different VRP variants. To address these issues, we take a novel perspective on model architecture in this study. Specifically, we propose a plug- and-play Entropy-based Scaling Factor (ESF) and a Distribution-Specific (DS) decoder to enhance the size and distribution generalization, respectively. ESF adjusts the attention weight pattern of the model towards familiar ones discovered during training when solving VRPs of varying sizes. The DS decoder explicitly models VRPs of multiple training distribution patterns through multiple auxiliary light decoders, expanding the model representation space to encompass a broader range of distributional scenarios. We conduct extensive experiments on both synthetic and widely recognized real-world benchmarking datasets and compare the performance with seven baseline models. The results demonstrate the effectiveness of using ESF and DS decoder to obtain a more generalizable model and showcase their applicability to solve different VRP variants, i.e., traveling salesman problem and capacitated VRP. Notably, our proposed generic components require minimal computational resources, and can be effortlessly integrated into conventional generalization strategies to further elevate model generalization.

关键词： Vehicle Routing Problem Generalization Model architecture neural combinatorial optimization

来源：评论

学校读者我要写书评

暂无评论

A Reinforcement Learning Method for Optical Thin-Film Design

引用

IEICE TRANSACTIONS ON ELECTRONICS 2022年第2期E105C卷 95-101页

作者： Jiang, Anqing Yoshie, Osamu Waseda Univ Grad Sch Informat Prod & Syst Kitakyushu 8080135 Japan

Machine learning, especially deep learning, is dramatically changing the methods associated with optical thin-film inverse design. The vast majority of this research has focused on the parameter optimization (layer thickness, and structure size) of optical thin-films. A challenging problem that arises is an automated material search. In this work, we propose a new end-to-end algorithm for optical thin-film inverse design. This method combines the ability of unsupervised learning, reinforcement learning and includes a genetic algorithm to design an optical thin-film without any human intervention. Furthermore, with several concrete examples, we have shown how one can use this technique to optimize the spectra of a multi-layer solar absorber device.

关键词： optical thin-film design reinforcement learning neural combinatorial optimization

来源：评论

学校读者我要写书评

暂无评论

A Reinforcement Learning Approach for Rebalancing Electric Vehicle Sharing Systems

引用

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 2022年第7期23卷 8704-8714页

作者： Bogyrbayeva, Aigerim Jang, Sungwook Shah, Ankit Jang, Young Jae Kwon, Changhyun Univ S Florida Dept Ind & Management Syst Engn Tampa FL 33620 USA Korea Adv Inst Sci & Technol Dept Ind & Syst Engn Daejeon 34141 South Korea

This paper proposes a reinforcement learning approach for nightly offline rebalancing operations in free-floating electric vehicle sharing systems (FFEVSS). Due to sparse demand in a network, FFEVSS requires relocation of electrical vehicles (EVs) to charging stations and demander nodes, which is typically done by a group of drivers. A shuttle is used to pick up and drop off drivers throughout the network. The objective of this study is to solve the shuttle routing problem to finish the rebalancing work in minimal time. We consider a reinforcement learning framework for the problem, in which a central controller determines the routing policies of a fleet of multiple shuttles. We deploy a policy gradient method for training recurrent neural networks and compare the obtained policy results with heuristic solutions. Our numerical studies show that unlike the existing solutions in the literature, the proposed methods allow solving the general version of the problem with no restrictions on the urban EV network structure and charging requirements of EVs. Moreover, the learned policies offer a wide range of flexibility, resulting in a significant reduction in the time needed to rebalance the network.

关键词： Shared mobility reinforcement learning neural combinatorial optimization vehicle routing

来源：评论

学校读者我要写书评

暂无评论

Simultaneous task and energy planning using deep reinforcement learning

引用

INFORMATION SCIENCES 2022年 607卷 931-946页

作者： Wang, Di Hu, Mengqi Weir, Jeffery D. Univ Illinois Dept Mech & Ind Engn Chicago IL 60607 USA Air Force Inst Technol Dept Operat Sci Wright Patterson AFB OH 45433 USA

To improve energy awareness of unmanned autonomous vehicles, it is critical to co optimize task planning and energy scheduling. To the best of our knowledge, most of the existing task planning algorithms either ignore energy constraints or make energy scheduling decisions based on simple rules. To bridge these research gaps, we propose a combinatorial optimization model for the simultaneous task and energy planning (STEP) problem. In this paper, we propose three variants of STEP problems (i) the vehicle can visit stationary charging stations multiple times at various locations;(ii) the vehicle can efficiently coordinate with mobile charging stations to achieve zero waiting time for recharging, and (iii) the vehicle can maximally harvest solar energy by considering time variance in solar irradiance. Besides, in order to obtain fast and reliable solutions to STEP problems, we propose a neural combinatorial optimizer using the deep reinforcement learning algorithm with a proposed link information filter. The near-optimal solutions can be obtained very fast without solving the problem from scratch when environments change. Our simulation results demonstrate that (i) our proposed neural optimizer can find solutions close to the optimum and outperform the exact and heuristic algorithms in terms of computational cost;(ii) the end-to-end learning (directly mapping from perceptions to control) model outperforms the traditional learning (mapping from perception to prediction to control) model. (c) 2022 Elsevier Inc. All rights reserved.

关键词： Simultaneous task and energy planning neural combinatorial optimization Deep reinforcement learning End-to-end learning Sequence-to-sequence decision

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：