Stochastic Simulation-based Machine Learning System for Duration and Cost Risk Evaluation of River Dredging Projects

• Main Authors: Ji-Wei Lin, 林季葦
• Other Authors: Jui-Sheng Chou
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/xm8787

碩士 === 國立臺灣科技大學 === 營建工程系 === 107 === The government has recently begun to value dredging engineering issues. The Water Resources Agency (WRA) seeks to maximize the effectiveness of various engineering projects. However, dredging engineering projects are often ineffective in reality, because dredging engineering has high uncertainty, and always involves many stakeholders. We found that many dredging project contractors submit tenders using construction period and cost estimation methods based on cases that have been undertaken on the last year. However, landform often changes owing to recent climate change, earthquakes, typhoons and other disasters. Therefore, only evaluating the projects from last year is inadequate, and often leads to increase in construction period and cost if the scope of the project is not clearly defined is or missing at the beginning of the construction.Based on previous research, this study aims to estimate the construction period and cost at the beginning of the construction. The first step is to collect the cloud data, outlay, revenue, acceptance certificate and other past data of dredging engineering from 2013 to 2018, and create the database. The next step is to consider the accessibility of quantitative variable data after completing the database, and after repeatedly discussing with experts and scholars.These are the sand, the gravel, the mud ratio, the average price of the muds, the total dredging amount and the total cost of the tender. Artificial intelligence technology is then adopted to build a deterministic model for dredging project duration and cost.