Productivity forecasting of newly added workers based on time-series analysis and site learning

Hyunsoo Kim, Hyun Soo Lee, Moonseo Park, Changbum R. Ahn, Sungjoo Hwang

Research output: Contribution to journalArticlepeer-review

4 Scopus citations


Adding new laborers during construction is usually considered the easiest option to execute when a schedule delay occurs in a construction project. However, determining the proper number of new laborers to add is quite challenging because newly added laborers' short-term productivity for their first several production cycles could be significantly different from that of existing laborers. While existing studies suggest that newly added laborers' site-learning may cause such a difference, this process has not been considered when forecasting newly added laborers' short-term productivity. In this context, this study presents a method that takes into account site-learning effects and the periodic characteristics of newly added laborers' short-term productivity. The periodic characteristics of productivity are analyzed based on a time-series model of existing laborers' productivity. Then, the impact of the site-learning effect on the productivity is considered based on existing learning-effect theory. An illustrative example demonstrates the accuracy and usefulness of the presented method. Its results indicate that the consideration of the site-learning effect prevents the frequent and counterproductive underestimation of the required number of newly added laborers in establishing an accelerated recovery schedule.

Original languageEnglish
Article number5015008
JournalJournal of Construction Engineering and Management
Issue number9
StatePublished - 1 Sep 2015

Bibliographical note

Publisher Copyright:
© 2015 American Society of Civil Engineers.


  • Productivity forecasting
  • Quantitative methods
  • Schedule delay
  • Site-learning effect
  • Time-series analysis


Dive into the research topics of 'Productivity forecasting of newly added workers based on time-series analysis and site learning'. Together they form a unique fingerprint.

Cite this