Joint Optimisation of ASC Network Structure and Product Variants Using Utility-Complexity Trade-Off = 효용-복잡도 트레이드오프를 고려한 ASC 네트워크 구조와 제품 다양성의 공동 최적화

다국어 초록 (Multilingual Abstract)

This research extends prior work on structural complexity and multi-objective optimisation in assembly supply chains by developing a mathematical model that identifies the optimal network configuration and product variety mix. The model incorporates both efficiency and complexity performance measures to capture the trade-off between customer satisfaction and operational manageability. In addition, we introduce a complexity-constrained version to compare and present a different approach. Through optimisation and sensitivity analysis, the study reveals how changes in network structure, production capabilities, and product design affect overall system performance and complexity

국문 초록 (Abstract)

본 연구는 조립공급망（Assembly Supply Chain, ASC）에서 제품 다양성과 네트워크 구조를 동시에 최적화하기 위한 수리적 모형을 제안함으로써 구조적 복잡성 및 다목적 최적화에 관한 기존 연구를 확장하였다. 제안된 모형은 고객 효용과 시스템 복잡성을 동시에 고려하여 고객 만족과 운영 효율성 간의 효용-복잡도 트레이드오프를 정량적으로 평가할 수 있도록 설계되었다. 또한 기존 효율성 중심의 최적화 모형과 비교하기 위하여 복잡도 제약 기반 최적화 모형을 함께 제안하였다. 다양한 최적화 실험과 민감도 분석을 수행한 결과, 네트워크 구조, 생산 능력 및 제품 설계의 변화가 전체 시스템의 성능과 복잡성에 중요한 영향을 미치는 것으로 나타났다. 본 연구는 제품 다양성과 공급망 구조를 통합적으로 고려하는 의사결정 프레임워크를 제시함으로써 학문적·실무적 시사점을 제공한다

목차 (Table of Contents)

• 1 Introduction 1
• 2 Literature Review 6
• 2.1 ASC Network Complexity 6
• 2.2 Product Variety and Operational Complexity 9
• 2.3 Integrated Optimisation of Utility and Complexity 11
• 2.4 Research Gap and Contribution Position 13
• 3 Model Framework 17
• 3.1 ASC Network Representation 17
• 3.2 Product and Module Variants 20
• 3.3 Consumer Choice Behaviour 23
• 3.4 Demand Propagation 26
• 3.5 System Complexity 29
• 3.5.1 Probability Definitions for Complexity Measurements 31
• 3.5.2 Assembly Line Complexity 33
• 3.5.3 Network Complexity 36
• 4 Enumeration of ASC Network Structures 39
• 4.1 Feasible Network Definition 40
• 4.2 Tree-Based Representation 41
• 4.3 Enumeration via Recursive Partitioning 43
• 4.4 Enumeration Algorithm 45
• 4.5 Consistency of Variant Mapping 47
• 4.6 Properties of Generated ASC Networks 48
• 5 Problem Formulation and Solution Approach 50
• 5.1 Sets 50
• 5.2 Parameters 51
• 5.3 Decision Variables 52
• 5.4 Relative Utility and Complexity 53
• 5.5 Objective Function 55
• 5.6 Total Optimisation Model 56
• 5.7 Alternative Formulation: Complexity-Constrained Model 60
• 5.8 Joint Network and Product Assortment Optimisation 62
• 5.9 Computational Strategy 64
• 5.10Solution Approach 66
• 5.11Illustrative Example 69
• 5.11.1Feasible ASC network structures 70
• 5.11.2Product selection patterns 71
• 5.11.3Influence of the complexity budget 75
• 5.11.4Network selection 77
• 6 Computational Experiments 81
• 6.1 Experimental Setup and Data Generation 82
• 6.1.1 Instance Construction 82
• 6.1.2 Network Generation and Classification 88
• 6.1.3 Complexity Evaluation 89
• 6.1.4 Optimisation Models 90
• 6.1.5 Solution Approach and GA Configuration 92
• 6.1.6 Experimental objectives 94
• 6.2 Interpretation of Performance Measures 95
• 6.3 Baseline Computational Results 98
• 6.3.1 Overall performance comparison 98
• 6.3.2 Utility-complexity trade-off across evaluated networks 101
• 6.3.3 Product variety and market utility 105
• 6.3.4 Product variety and system complexity 107
• 6.3.5 Network topology selection 109
• 6.3.6 Assortment similarity between formulations 113
• 6.3.7 Mechanisms underlying model and topology differences 116
• 6.4 Sensitivity Analysis 119
• 6.4.1 Sensitivity to the complexity bound 120
• 6.4.2 Sensitivity to complexity component weights 123
• 6.4.3 Solution stability under changes in the complexity bound 127
• 6.4.4 Solution stability under changes in complexity component weights 131
• 6.4.5 Summary of sensitivity findings 136
• 7 Conclusions and Discussion 138
• References 145
• A Optimisation Algorithms 148
• B Generation of Computational Instances 149
• C ASC Network Enumeration Procedure 151
• D Ratio-Based Optimisation Procedure 153
• E Genetic Algorithm Settings 154
• F Complexity-Constrained Optimisation Procedure 156
• G Network Topology Classification 158
• H Complexity Decomposition 159
• I Experimental Design 160
• J Sensitivity Analysis 161
• K Topology Mechanism Analysis 162
• L Representative Output Measures 164
• M Additional Figures and Tables 165
• M.1 Experimental Settings 166
• M.2 Stability of the Constrained Formulation 169
• M.3 Changes in Product Assortments Across Complexity Bounds 171
• M.4 Changes in Product Assortments Across Weight Settings 173
• M.5 Utility Characteristics of Modified Products 176
• M.6 Additional Product Selection Analyses 176
• N Implementation Environment 176
• 국문초록 181