Assessing Project Staffing Requirements Using Unsupervised Clustering Techniques

Arthur W.T. Leung1 & C.M. Tam2

1Senior Lecturer, Division of Building Science and Technology, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong, Phone: (852) 27887617, Fax:(852)27889716, bsawtl@cityu.edu.hk
2Professor, Department of Building & Construction, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong, Phone: (852) 27887620, Fax:(852)27887612, bctam@cityu.edu.hk

Abstract

Standardization, prefabrication and mechanization have been adopted, with similar concepts of lean construction, by the Hong Kong SAR government to solve housing problems since 1980. Project managers usually believe that the number of supervisory staff is directly proportional to the scale of a project. However, it is observed that there are significant variations in the allocation of supervisory staff. The objective of this pilot study attempts to explore the relationship between the scale of building projects, in terms of project clusters, and staffing strategies. In order to derive objective classification for building projects, objective project data has been used for forming project clusters using the Self-organizing Map (SOM) algorithm, which is a well-known unsupervised clustering technique. The project clusters formed represent the natural grouping in terms of similarity. The relationships between supervisory staffing patterns and project clusters have been reviewed. The findings identifies that there is a significant difference between staffing strategies for standardized public housing project and supervisory staff has been reduced proportionally to the project scale. The results shed some light on the understanding of staffing practice adopted by contractors in Hong Kong and provide some insight for future research.

Keywords

project cluster, project scale, staffing cost, unsupervised clustering technique, site organization

Files

Reference

Leung, A.W. & Tam, C. 2008, 'Assessing Project Staffing Requirements Using Unsupervised Clustering Techniques' In:, Tzortzopoulos, P. & Kagioglou, M., 16th Annual Conference of the International Group for Lean Construction. Manchester, UK, 16-18 Jul 2008. pp 613-622

Download: BibTeX | RIS Format