Deep learning based high accuracy heuristic approach for knapsack interdiction problem

IF 4.1 2区 工程技术 Q2 COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS
Sunhyeon Kwon, Hwayong Choi, Sungsoo Park
{"title":"Deep learning based high accuracy heuristic approach for knapsack interdiction problem","authors":"Sunhyeon Kwon,&nbsp;Hwayong Choi,&nbsp;Sungsoo Park","doi":"10.1016/j.cor.2024.106965","DOIUrl":null,"url":null,"abstract":"<div><div>Interdiction problems are a subfamily of bilevel optimization problems, characterized by a hierarchical structure involving two agents: a leader and a follower. In these problems, the objective functions of the leader and the follower are identical but are optimized in opposite directions. In this paper, we focus on the knapsack interdiction problem, where the leader and the follower compete for a shared set of items. While exact algorithms exist to solve this problem, they may not be suitable for slightly larger instances. As an alternative to exact algorithms, we propose a heuristic approach based on deep learning. Our method involves training three types of neural networks: a core network that aggregates information about the problem, a classification network that directly identifies solutions, and an identification network that assesses the reliability of the classification network’s results. Our algorithm successfully finds optimal or near-optimal solutions up to 21 times faster than the exact algorithm for both the training data sizes and larger problem instances.</div></div>","PeriodicalId":10542,"journal":{"name":"Computers & Operations Research","volume":"176 ","pages":"Article 106965"},"PeriodicalIF":4.1000,"publicationDate":"2024-12-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computers & Operations Research","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0305054824004374","RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"COMPUTER SCIENCE, INTERDISCIPLINARY APPLICATIONS","Score":null,"Total":0}
引用次数: 0

Abstract

Interdiction problems are a subfamily of bilevel optimization problems, characterized by a hierarchical structure involving two agents: a leader and a follower. In these problems, the objective functions of the leader and the follower are identical but are optimized in opposite directions. In this paper, we focus on the knapsack interdiction problem, where the leader and the follower compete for a shared set of items. While exact algorithms exist to solve this problem, they may not be suitable for slightly larger instances. As an alternative to exact algorithms, we propose a heuristic approach based on deep learning. Our method involves training three types of neural networks: a core network that aggregates information about the problem, a classification network that directly identifies solutions, and an identification network that assesses the reliability of the classification network’s results. Our algorithm successfully finds optimal or near-optimal solutions up to 21 times faster than the exact algorithm for both the training data sizes and larger problem instances.
基于深度学习的背包拦截问题高精度启发式方法
阻断问题是双层优化问题的一个子集,其特征是涉及两个主体:领导者和追随者的分层结构。在这些问题中,领导者和追随者的目标函数是相同的,但在相反的方向上进行优化。本文研究了领导者和追随者争夺一组共享物品的背包拦截问题。虽然存在精确的算法来解决这个问题,但它们可能不适合稍微大一点的实例。作为精确算法的替代方案,我们提出了一种基于深度学习的启发式方法。我们的方法涉及训练三种类型的神经网络:聚合问题信息的核心网络,直接识别解决方案的分类网络,以及评估分类网络结果可靠性的识别网络。对于训练数据大小和更大的问题实例,我们的算法成功地找到最优或接近最优的解决方案,速度比精确算法快21倍。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
Computers & Operations Research
Computers & Operations Research 工程技术-工程:工业
CiteScore
8.60
自引率
8.70%
发文量
292
审稿时长
8.5 months
期刊介绍: Operations research and computers meet in a large number of scientific fields, many of which are of vital current concern to our troubled society. These include, among others, ecology, transportation, safety, reliability, urban planning, economics, inventory control, investment strategy and logistics (including reverse logistics). Computers & Operations Research provides an international forum for the application of computers and operations research techniques to problems in these and related fields.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信