Automate Construction Scheduling at the Pre-construction Stage – A Case Study
Accepted version
Peer-reviewed
Repository URI
Repository DOI
Change log
Authors
Abstract
Construction schedules are the timetables showing the planned start and finish dates for activities within projects, and thus serve critical roles to instruct construction execution and monitor project progress. However, the construction industry has been suffering from delays and cost overrun for decades. It’s becoming important to produce quality schedules with accurate contingency allocation and risk mitigation plans. Experienced schedulers produce quality schedules based on personal experiences and knowledge. However, such tacit knowledge hasn’t been captured, stored and shared with inexperienced schedulers. This paper proposed a Graph-based Schedule Mining (GSM) method to capture, store and reuse the tacit knowledge in the construction schedules. The GSM method consists of three stages: feature generation, data annotation, and classification. The GSM method is implemented on a bridge project. We then compare with the original schedule regarding the time- and cost-efficient. The results indicated that the time-productivity of identified earthwork and foundation construction sequences are 8.7% and 15.5% closer to the actual construction than planned. The GSM method can be implemented at both project and organisational levels to help schedulers to initiate more accurate schedules quicker and organisations to establish knowledge management systems with less labour investment.