Publications

Detailed Information

Integer programming models and exact algorithms for the two-dimensional two-staged knapsack problem : 2차원 2단계 배낭문제에 대한 정수계획모형 및 최적해법

DC Field Value Language
dc.contributor.advisor이경식-
dc.contributor.author강수호-
dc.date.accessioned2021-11-30T02:09:22Z-
dc.date.available2021-11-30T02:09:22Z-
dc.date.issued2021-02-
dc.identifier.other000000164006-
dc.identifier.urihttps://hdl.handle.net/10371/175194-
dc.identifier.urihttps://dcollection.snu.ac.kr/common/orgView/000000164006ko_KR
dc.description학위논문 (석사) -- 서울대학교 대학원 : 공과대학 산업공학과, 2021. 2. 이경식.-
dc.description.abstractIn this thesis, we study integer programming models and exact algorithms for the two-dimensional two-staged knapsack problems, which maximizes the profit by cutting a single rectangular plate into smaller rectangular items by two-staged guillotine cuts. We first introduce various integer programming models, including the strip-packing model, the staged-pattern model, the level-packing model, and the arc-flow model for the problem. Then, a hierarchy of the strength of the upper bounds provided by the LP-relaxations of the models is established based on theoretical analysis. We also show that there exists a polynomial-size model that has not been proven yet as far as we know. Exact methods, including branch-and-price algorithms using the strip-packing model and the staged-pattern model, are also devised. Computational experiments on benchmark instances are conducted to examine the strength of upper bounds obtained by the LP-relaxations of the models and evaluate the performance of exact methods. The results show that the staged-pattern model gives a competitive theoretical and computational performance.-
dc.description.abstract본 논문은 2단계 길로틴 절단(two-staged guillotine cut)을 사용하여 이윤을 최대화하는 2차원 2단계 배낭 문제(two-dimensional two-staged knapsack problem: 이하 2TDK)에 대한 정수최적화 모형과 최적해법을 다룬다. 우선, 본 연구에서는 스트립패킹모형, 단계패턴모형, 레벨패킹모형, 그리고 호-흐름모형과 같은 정수최적화 모형들을 소개한다. 그 뒤, 각각의 모형의 선형계획완화문제에 대해 상한강도를 이론적으로 분석하여 상한강도 관점에서 모형들 간 위계를 정립한다. 또한, 본 연구에서는 2TDK의 다항크기(polynomial-size) 모형의 존재성을 처음으로 증명한다. 다음으로 본 연구는 2TDK의 최적해를 구하는 알고리즘으로써 패턴기반모형들에 대한 분지평가 알고리즘과 레벨패킹모형을 기반으로 한 분지절단 알고리즘을 제안한다. 단계패턴모형이 이론적으로도 가장 좋은 상한강도를 보장할 뿐만 아니라, 계산 분석을 통해 단계패턴모형을 기반으로 한 분지평가 알고리즘이 제한된 시간 내 좋은 품질의 가능해를 찾음을 확인하였다.-
dc.description.tableofcontentsAbstract i
Contents iv
List of Tables vi
List of Figures vii
Chapter 1 Introduction 1
1.1 Problem Description . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Organization of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . 8
Chapter 2 Integer Programming Models for 2TDK 9
2.1 Pattern-based Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Arc-flow Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Level Packing Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Chapter 3 Theoretical Analysis of Integer Programming Models 20
3.1 Upper Bounds of AF and SM(1;1) . . . . . . . . . . . . . . . . . . 20
3.2 Upper Bounds of ML, PM(d), and SM(d; d) . . . . . . . . . . . . . . 21
3.3 Polynomial-size Model . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Chapter 4 Exact Methods 33
4.1 Branch-and-price Algorithm for the Strip Packing Model . . . . . . . 34
4.2 Branch-and-price Algorithm for the Staged-pattern Model . . . . . . 39
4.2.1 The Standard Scheme . . . . . . . . . . . . . . . . . . . . . . 39
4.2.2 The Height-aggregated Scheme . . . . . . . . . . . . . . . . . 40
4.3 Branch-and-cut Algorithm for the Modified Level Packing Model . . 44
Chapter 5 Computational Experiments 46
5.1 Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.2 Upper Bounds Comparison . . . . . . . . . . . . . . . . . . . . . . . 49
5.2.1 A Group of Small Instances . . . . . . . . . . . . . . . . . . . 49
5.2.2 A Group of Large Instances . . . . . . . . . . . . . . . . . . . 55
5.2.3 Class 5 Instances . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.3 Solving Instances to Optimality . . . . . . . . . . . . . . . . . . . . . 65
5.3.1 A Group of Small Instances . . . . . . . . . . . . . . . . . . . 65
5.3.2 A Group of Large Instances . . . . . . . . . . . . . . . . . . . 69
5.3.3 Class 5 Instances . . . . . . . . . . . . . . . . . . . . . . . . . 72
Chapter 6 Conclusion 77
Bibliography 79
국문초록 83
-
dc.format.extentvii, 83-
dc.language.isoeng-
dc.publisher서울대학교 대학원-
dc.subjectInteger Programming-
dc.subjectTwo-dimensional two-staged knapsack problem-
dc.subjectDantzig-Wolfe decomposition-
dc.subjectBranch-and-price algorithm-
dc.subject정수최적화 모형-
dc.subject2차원 2단계 배낭 문제-
dc.subject분지평가 알고리즘-
dc.subject분지절단 알고 리즘-
dc.subject비교분석-
dc.subject.ddc670.42-
dc.titleInteger programming models and exact algorithms for the two-dimensional two-staged knapsack problem-
dc.title.alternative2차원 2단계 배낭문제에 대한 정수계획모형 및 최적해법-
dc.typeThesis-
dc.typeDissertation-
dc.contributor.AlternativeAuthorSuho Kang-
dc.contributor.department공과대학 산업공학과-
dc.description.degreeMaster-
dc.date.awarded2021-02-
dc.identifier.uciI804:11032-000000164006-
dc.identifier.holdings000000000044▲000000000050▲000000164006▲-
Appears in Collections:
Files in This Item:

Altmetrics

Item View & Download Count

  • mendeley

Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.

Share