Artificial Intelligence-Based Inference Support Models for Construction Engineering and Management

• Main Authors: Minh-Tu Cao, 高明秀
• Other Authors: Min-Yuan Cheng
• Format: Others
• Language: en_US
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/77048761614268336370

博士 === 國立臺灣科技大學 === 營建工程系 === 103 === Problems in the field of construction management are sophisticated, highly uncertain, and context-dependent. Therefore, using artificial intelligence (AI) to tackle these problems is a promising direction for research. The present research integrates Multivariate Adaptive Regression Splines (MARS), Radial Basis Function Neural Network (RBFNN), Artificial Bee Colony (ABC), and Fuzzy Logic (FL) to develop three novel inference models. The first model, the Evolutionary Multivariate Adaptive Regression Splines (EMARS), incorporates MARS and ABC. The EMARS resolves regression problems using input values for the underlying function mapping response and the various factors of influence that are provided by civil engineers. The second model, the Self-adaptive Structure Radial Basis Function Inference Model (SSRIM), fuses MARS, RBFNN, and ABC. The SSRIM is an efficient model for addressing inference tasks that have many un-assessed potential factors of influence. After the MARS removes the redundant (neutral) factors from the set of input variables, the ABC-optimized RBFNN uses the remaining input variables to execute the supervised learning task. Finally, the Intelligent Fuzzy Radial Basis Function Neural Network Inference Model (IFRIM), hybridizes RBFNN, FL, and ABC. In the IFRIM, FL handles vague input information, RBFNN handles the fuzzy input-output mapping relationships, and the ABC search engine employs optimization to identify the most suitable tuning parameters for RBFNN and FL based on minimal error estimation. Experimental results obtained from the applications of these newly established AI models demonstrate that these models may significantly enhance the ability of decision makers to resolve problems in the field of construction engineering and management.