vii TABLE OF CONTENTS CHAPTER 1 2 TITLE PAGE DECLARATION DEDICATION ACKNOWLEDGEMENT ii iii iv ABSTRACT ABSTRAK TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES v vi vii x xi LIST OF ABBREVIATIONS LIST OF SYMBOLS LIST OF APPENDICES xiii xiv xvi INTRODUCTION 1.1 Introduction 1.2 Research background 1 1 2 1.3 1.4 1.2.1 Job Shop Problem (JSP) 1.2.2 Ant Colony Optimization (ACO) Problem Statement Objective of the Study 2 3 5 5 1.5 1.6 1.7 Scope of the Study Significance of the Study Thesis Organization 6 6 7 LITERATURE REVIEW 2.1 Introduction 2.2 Behaviour of Real Ants 2.3 2.4 2.5 Classical Job Shop Problem Computational Complexity Shifting Bottleneck 8 8 8 11 13 13 viii 2.6 2.7 2.8 3 Review Relevant Research Research Findings on ACO Summary RESEARCH METHODOLOGY 3.1 Introduction 19 19 3.2 Job Shop Problem 3.2.1 Introduction 3.2.2 Definition for Job Shop 3.2.3 JSP Mathematical Formulation 19 19 20 20 3.2.4 3.2.5 3.2.6 3.2.7 3.2.8 22 23 24 26 27 3.3 Constraint for the JSP Assumptions for the JSP The JSP Representation Problem Formulation Graph Representation for Job Shop Problem The Solution Approaches 3.3.1 Methodology of Ant Colony Optimization 3.3.2 The new ACO-based Algorithm 3.3.2.1 Initialization 3.3.2.2 3.3.2.3 3.3.2.4 3.3.2.5 3.4 4 5 14 17 18 The Ant Architecture Daemon Action Pheromone Update Terminating Condition Summary 30 32 34 36 36 38 38 40 40 INDUSTRIAL PROBLEM EXPLORATION 41 4.1 4.2 4.3 4.4 Introduction Introduction to Manufacturing Company The Manufacturing Framework Company Problem Description 41 41 42 47 4.5 4.6 4.7 The Case Study Sensitivity Analysis Summary 51 59 67 IMPLEMENTATION OF ANT COLONY OPTIMIZATION FOR JOB SHOP PROBLEM 68 5.1 Introduction 68 5.2 Development of ACO Algorithm 68 ix 5.3 5.4 5.5 6 7 Algorithm for ACO Data Diagram Summary 70 84 89 SYSTEM DEVELOPMENT FOR ACO SOFTWARE 6.1 Introduction 90 90 6.2 6.3 6.4 6.5 Program Fundamentals Programming with Microsoft Visual Studio The Visual Studio Application Program Visualization 90 91 92 93 6.6 Summary 104 ANALYSIS OF RESULTS, CONCLUSION AND RECOMMENDATION 7.1 Introduction 7.2 Results 7.3 Analyze of The Results 105 105 105 109 7.4 7.5 7.6 113 114 114 REFERENCES Appendices A – 7 Conclusion Contribution Recommendation for Future Research 115 ?? – ?? x LIST OF TABLES TABLE NO. TITLE PAGE 2.1 2.2 2.3 The Instance data with 3 machines and 3 jobs. Relevant Research on Ant Colony Optimization . Relevant Research on Job Shop Problem. 12 15 16 2.4 Relevant Research on Job Shop Scheduling Problem in Ant Colony Optimization. 17 3.1 The Processing Time for 2/3/G/Cmax job shop problem. 29 4.1 4.2 The Processing Time for 2/3/G/Cmax job shop problem. Random values of Cmax and makespan depending on the 57 parameter values of α, β and ρ. Random values of Cmax and makespan depend on the values of β parameters, when ρ = 0 . Random value of Cmax and makespan depending on the values 60 of β parameters, when ρ = 1 and α = 0. Random values of Cmax and makespan depend on the values of β parameters, when ρ = 2. Random values of Cmax and makespan depend on the values of β parameters, when ρ = 1 and α = 1. 62 4.3 4.4 4.5 4.6 4.7 4.8 Random values of Cmax and makespan depend on the values of ρ parameters, when β = 0. Random values of Cmax and makespan depend on the values of ρ parameters, when β = 1. 5.1 Short caption 7.1 7.2 7.3 The Stimulation Results. Results for makespan with 6 machines. Comparison the results between Shifting Bottleneck Heuristic Method [1] with the new ACO-based Algorithm. 61 63 64 65 66 85 110 110 112 xi LIST OF FIGURES FIGURE NO. TITLE PAGE 2.1 2.2 2.3 Ant experimental for the bridge experiment. The ants are moving on the straight line. An unexpected obstacle has interrupted the initial path. 9 9 10 2.4 2.5 The same situation on the other side of the obstacle. Visual of ants choosing the shorter path. 10 11 3.1 3.2 3.3 The definition of a 2/3/G/Cmax job shop problem into graph. The result for a 2/3/G/Cmax job shop problem into graph. The flow chart for work explanation. 28 30 32 3.4 The problem solving step. 33 4.1 4.2 4.3 The framework for the warehouse process. The typical job shop layout problem that forms the case study. The data for the product customized E07 rack. 43 44 48 4.4 4.5 4.6 4.7 4.8 The processing time for the product. The data input with 6 machine. The graph for input data Power Press proses. The graph for input data Tapping process. The graph for input data Clinching Process. 49 50 52 53 54 4.9 4.10 4.11 The graph for input data Bending Process. The graph for input data Machining Process. The total number of edges versus total number of jobs run on 6 machines. 55 56 4.12 57 4.13 4.14 4.15 The total number of edges versus total number of nodes and operations. The graph on sensitivity analysis for Table 3. The graph for Sensitivity analysis for Table 4.4. The graph for sensitivity analysis for Table 5.5. 58 61 62 63 4.16 4.17 The graph for sensitivity analysis for Table 6. The graph for sensitivity analysis for Table 7. 64 65 xii 4.18 The graph for sensitivity analysis for Table 4.8. 66 5.1 5.2 5.3 The Pseudocode of ACO algorithm. The framework for the proposed ACO algorithm. The Gantt Diagram for the Power Press Process. 68 70 86 5.4 5.5 5.6 5.7 The Gantt Diagram for the Tapping Process. The Gantt Diagram for the Clinching Process. The Gantt Diagram for the Bending Process. The Gantt Diagram for the Machining Process. 86 87 88 88 6.1 6.2 6.3 6.4 The Welcome GUI for the program. The Project Properties for the Case Study. The Input Data for the CNC Machining. The Input Data for the Bending process. 93 94 95 96 6.5 6.6 6.7 6.8 6.9 The Input Data for the Clinching process. The Input Data for the Tapping process. The Input Data for the Power Press process. The Results for the case study. The Debug for the case study. 97 98 99 100 101 6.10 The flow run for the software development. 103 7.1 7.2 7.3 The consequence for Power Press Process. The Consequence for Tapping Process. The Consequence for Clinching Process. 106 106 107 7.4 7.5 The consequence for Bending Process. The consequence for CNC Machining Process. 108 109 xiii LIST OF ABBREVIATIONS ACO – Ant Colony Optimization JSP – Job Shop Problem Np-hard – Non-deterministic Polynomial Time Hard TSP – Traveling Salesman Problem – xiv LIST OF SYMBOLS Cmax – Makespan / total completion time m – Total of machine n – Total of job i – node i j – node j Mm – m Machine Jn – n Job Jj – Jobs at node j Oij – Operation at node i and node j Pij – Processing Time s – Source Node t – Sink Node λ – Wavelength α – Real positive parameters β – Real positive parameters ρ – Evaporation rate γ – Decision variable, generates a sequence |O| – Operations ηij – Heuristic value associates with the common Cij τij – Pheromone value at node i and j dij – Heuristic distance between node i and j G – A representative of a graph Sije – Empty Solution xv Sg – Good solution Lk – Tour length of the k-th ant – xvi LIST OF APPENDICES APPENDIX TITLE PAGE