Using the Fast Messy Genetic Algorithm and the Simulation Mechanism to Develop the Optimal Process for Design Project

• Main Authors: Su-Ting Hsu, 許淑婷
• Other Authors: Chung-Wei Feng
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/60168947694989187902

碩士 === 國立成功大學 === 土木工程學系碩博士班 === 94 === The design project is typically involved with processing a large amount of information from various participants. However, owing to the uncertainty within the information, engineers have to repeatedly perform design activities, such as exchanging information with other participants or processing information, to produce an appropriate result. This persistent information exchanging process may involve many activities and come into one or several iterative design loops. As there are more activities involving within the iteration loop, the productivity of the design project could fall down easily. Furthermore, because the relationships among the design activities are mainly determined by information flows, the sequence of performing design activities should be properly planned to avoid unnecessary iteration loops. Therefore, it is essential to develop a model that can improve the productivity of the design phase by reducing the unnecessary iterations loops. Dependency structure matrix (DSM) and simulation are used methods to describe the design process in this research. By employing DSM, the relationships and the sequence of the design activities can be identified. Many researchers had developed various models to optimize the design process according to DSM or simulation method. However, the impact of information processing or communication within the indifferent design process is not thoroughly considered in these studies. In addition, as the number of activities within DSM increases, the optimization process becomes computationally inefficient. Therefore, it is necessary to build a more efficient model which not only thoroughly consider the impact of iteration loops but also find the optimal design process for the large-size design project. In this research, a simulation mechanism is applied to evaluating the design process. Furthermore, fast messy genetic algorithm (fmGA) is employed along with the simulation mechanism to build the optimization model of the design process. Results showed that, based on the proposed model, the impact of the iteration loops can be clearly identified and reduced. In addition, engineers can find the optimal design process for the large-scale design project efficiently with the computer implementation.