| Summary: | 碩士 === 國立成功大學 === 土木工程學系 === 89 === Abstract
On construction sites, construction engineers routinely face the problem of allocating temporary site-level facilities such as job-site offices, warehouses, and assorted workshops etc. to appropriate locations. With proper allocation of facilities, construction sites can be more efficient in term of production cost and time. Previous research to solve the facility layout problem focus on generating possible alternatives and selecting the best layout, which can be categorized into layout improvement, entire layout design, and partial layout design approaches. However, those approaches are rarely accepted in the construction site because of the complexity involved within the problem and their simplified assumptions. Recently, artificial intelligent-based approaches such as neural networks and genetic algorithms have been employed to clear up the complexity within the construction layout problem. However, it is still hard to apply these intelligent-based facility layout models to the practice field of construction because of the simplified assumptions they hold.
This research present an investigation of current practices on solving the construction facility layout problem and a new construction facility layout model that is more readily applicable to construction sites. This new construction facility layout model allows construction engineers to determine several parameters, such as size, different types of shapes, entrances and exit of facilities and construction site, which is more realistic in the construction sites. The workflows between facilities are also considered, while minimizing the total traveling distance between facilities. Along with this new facility layout model, a set of self-adapted algorithms that is based on the principles of Genetic Algorithms is developed. This set of algorithms employ the specific coding of facility shapes provides a more efficient tool to solve the construction layout problems. The results show that this new genetic algorithm can quickly generates the best solution. Furthermore, the new genetic algorithm, a user-friendly program which is integrated Microsoft Excel with C++ coding is built to provide a more practical tool for construction engineer. This program presents a new way of planning construction facility layout in a more realistic manner by exploring what-if scenarios.
|
|---|
