研究生: |
Pratama Mahardika Firdausi Pratama - Mahardika Firdausi |
---|---|
論文名稱: |
High Performance Concrete Compressive Strength Prediction Using Genetic Weighted Pyramid Operation Tree (GWPOT) High Performance Concrete Compressive Strength Prediction Using Genetic Weighted Pyramid Operation Tree (GWPOT) |
指導教授: |
鄭明淵
Min-Yuan Cheng |
口試委員: |
陳鴻銘
Hung-Ming Chen 陳介豪 Jieh-Haur Chen 曾仁杰 Ren-Jye Dzeng |
學位類別: |
碩士 Master |
系所名稱: |
工程學院 - 營建工程系 Department of Civil and Construction Engineering |
論文出版年: | 2013 |
畢業學年度: | 101 |
語文別: | 英文 |
論文頁數: | 131 |
中文關鍵詞: | Prediction 、Concrete Strength 、Genetic Algorithm 、Operation Tree 、Weighted Pyramid Operation Tree |
外文關鍵詞: | Prediction, Concrete Strength, Genetic Algorithm, Operation Tree, Weighted Pyramid Operation Tree |
相關次數: | 點閱:288 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
This study introduces the Genetic Weighted Pyramid Operation Tree (GWPOT) to build a formula to solve the problem of predicting high-performance concrete compressive strength. GWPOT is a new improvement of the genetic operation tree that consists of Genetic Algorithm, Weighted Operation Structure, and Pyramid Operation Tree. Model accuracy was compared against other widely used artificial intelligence (AI) models such as the Artificial Neural Network (ANN), Support Vector Machine (SVM), and Evolutionary Support Vector Machine Inference Model (ESIM). Results demonstrate GWPOT as an efficient approach that performs better than others competing models. Moreover, GWPOT can generate explicit formulas, which is an important advantage in practical applications while other black-box prediction techniques unable to generate explicit formulas.
This study introduces the Genetic Weighted Pyramid Operation Tree (GWPOT) to build a formula to solve the problem of predicting high-performance concrete compressive strength. GWPOT is a new improvement of the genetic operation tree that consists of Genetic Algorithm, Weighted Operation Structure, and Pyramid Operation Tree. Model accuracy was compared against other widely used artificial intelligence (AI) models such as the Artificial Neural Network (ANN), Support Vector Machine (SVM), and Evolutionary Support Vector Machine Inference Model (ESIM). Results demonstrate GWPOT as an efficient approach that performs better than others competing models. Moreover, GWPOT can generate explicit formulas, which is an important advantage in practical applications while other black-box prediction techniques unable to generate explicit formulas.
ABSTRACT i
ACKNOWLEDGEMENT ii
TABLE OF CONTENTS iv
LIST OF FIGURES vi
LIST OF TABLES viii
ABBREVIATIONS x
CHAPTER 1: INTRODUCTION 1
1.1 Study Motivation 1
1.2 Study Objective 3
1.3 Study Scope 3
1.4 Study Methodology 3
1.5 Study Outline 6
CHAPTER 2: LITERATURE REVIEW 7
2.1 High Performance Concrete (HPC) 7
2.2 Operation Tree (OT) 8
2.3 Genetic Algorithm (GA) 10
2.4 Weighted Operation Structure (WOS) 15
CHAPTER 3: GENETIC WEIGHTED PYRAMID OPERATION TREE (GWPOT) 17
3.1 GWPOT Architecture 17
3.2 Numerical Example 20
3.3 Modified Predicted Output Value 25
CHAPTER 4: CASE STUDY 27
4.1 Dataset 27
4.2 Tuning Parameter 27
4.3 k-Fold Cross Validation 28
4.4 Performance Measurement 29
4.5 Result and Discussion 31
4.6 Comparison 42
4.7 Sensitivity Analysis 47
CHAPTER 5: CONCLUSION AND RECOMMENDATIONS 52
5.1 Review Study and Purpose 52
5.2 Conclusions 52
5.3 Future Research Recommendations 53
REFERENCES 54
APPENDIX A: Historical Data 57
APPENDIX B: MatLab Source Code 82