Automated Construction Work Package Decision-Making Model Using RSMeans and LSTM Based Deep Learning
Abstract
Construction projects involve complicated workflows with efficient resource management. One of the methods for successful project planning is Work Packaging (WP.) Composition of the WP has been done by human understanding, resulting in several critical problems, such as challenging WP of the inexperienced, format and term inconsistency, and time-consuming and error-prone nature of the human activity. Thus, the WP assistant system is in need to solve the problems. This study aims to develop an automated WP decision-making prototype from detail section drawings. The research objectives are as follows: 1) Organize detail section drawings using RSMeans Assemblies Costs standard, 2) Construct prototype which automates WP decision procedure, 3) Evaluate the sensitivity of the decision output. This prototype produced about 95.24% testing accuracy from 314 datasets, significantly accurate for automated WP decisions. This research is expected to solve problems from human WP compositions and eventually contribute to the efficient WP organization of construction project entities.
Subject
Construction Work PackageRSMeans
Artificial Intelligence
Deep Learning
Automation
Automated Decision Making
Citation
Oh, John Seok (2022). Automated Construction Work Package Decision-Making Model Using RSMeans and LSTM Based Deep Learning. Master's thesis, Texas A&M University. Available electronically from https : / /hdl .handle .net /1969 .1 /197225.