Predicting Productivity Loss Caused by Change Orders Using Evolutionary Fuzzy Support Vector Machine Inference Model

• Main Authors: Dedy Kurniawan Wibowo, 容德慶
• Other Authors: Min-Yuan Cheng
• Format: Others
• Language: en_US
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/9a2dc3

碩士 === 國立臺灣科技大學 === 營建工程系 === 100 === Change orders in construction projects are very common and result in many negative impacts. The impact of change orders on labor productivity is difficult to quantify. A complex input-output relationship that measures the effect of change orders cannot be calculated using a traditional approach. In this study, Evolutionary Fuzzy Support Vector Machines Inference Model (EFSIM), which combines fuzzy logic (FL), support vector machine (SVM), and fast messy genetic algorithm (fmGA) is developed as a tool for predicting productivity loss caused by change orders. The SVM is utilized as a supervised learning technique for solving classification and regression problems. The advantages of FL in reckoning vagueness and uncertainty are exploited. Furthermore, fmGA is applied to optimize the model’s parameters. A case study regarding productivity loss caused by change orders is presented to demonstrate and to validate the performance of the proposed prediction model. Simulation results demonstrate EFSIM’s ability to predict the impact of change orders is outperformed compared to artificial neural network (ANN), support vector machine (SVM) and evolutionary support vector machine inference model (ESIM). Validation with previous studies shows that EFSIM successfully improve the accuracy and reliability of the prediction model.