碩士 / 國立臺灣科技大學 / 自動化及控制研究所 / 103 / Submarine pipeline is the main access route for transporting imported crude oil from the offshore oil dump station to the oil tank near the shore. In this study, the particle swarm optimization (PSO) and the self-learning particle swarm optimization (SLPSO) were used to obtain the optimal submarine pipeline layout planning considering the facts including the changes in the subsea terrain, pipeline flow, and pipeline lengths.
PSO is one of the most efficient tools for solving global optimization. Literatures have shown that when particles in the PSO make a subsequent selection, the same strategy is used. That is, only one learning model is used. Under this situation, due to social and cognitive model constraints, particles are likely to fall into local optimal solutions and thus be unable to process excessively complex problems. In order to resolve this problem, the self-learning algorithm was applied in this study. This algorithm divides learning strategies into four types to adapt to different environments, thus enabling each particle to choose a corresponding learning strategy based on their varied adaptation values.
In general, the best solution for engineering applications related to optimum path planning, i.e., the shortest path. In other words, it refers to the path requiring the lowest cost. However, in actual engineering applications, the degree of engineering difficulty and costs that arise based on different terrain conditions should be taken into consideration. Based on the viewpoint in this study, the shortest path does not represent the path with the lowest costs. Hence, the shortest path does not necessarily represent the path with the lowest cost. The SLPSO with different construction cost weights was applied in this thesis to solve submarine pipeline layout planning. The simulation results in this study show that the SLPSO can derive the optimal solution more efficiently and more accurately compared to the PSO.
| Identifer | oai:union.ndltd.org:TW/103NTUS5146026 |
| Date | January 2015 |
| Creators | Hwien-Wei Chen, 陳憲為 |
| Contributors | Sheng-Dong Xu, 徐勝均 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 70 |
Page generated in 0.0132 seconds