Semi-Automated Methods for Measuring Practice Conformance for Capital Projects

Abstract

The goal of this thesis is to explore semi-automated methods for measuring practice conformance for capital projects. Thorough measurement of practice conformance for capital projects typically requires manual audits. Surveys that may assist can often be subjective, non-repeatable and unverifiable, since they are self-reported. However, some of the tasks assigned to auditors are also non-repeatable, and they may be costly, time-consuming, tedious, and error-prone. Tools for assisting practice conformance measurements are in high demand in the construction domain. In response, various information technology-based and web deployed Benchmarking and Metrics (BM&M) programs have been introduced to reduce time and costs, to assist in providing repeatable and accurate results, and to increase efficiency and productivity of reporters and auditors. Moreover, moves toward automated practice conformance measurement are expected to reduce time and cost. Past studies have also resulted in significant advances in data mining, natural language processing, machine learning, computer vision and other artificial intelligence-based approaches toward complete automation, but technical limitations exist that constrain complete automation or make it impractical. An approach is needed to support practical, net beneficial, incremental steps toward automation of practice conformance measurement for capital projects that would assist capital project participants to improve project performance over time. To address this need, a new approach is proposed in this thesis. Additionally, a framework to beneficially increase automation is presented. Toolsets are explored that may make practice conformance measurement cheaper, faster, easier, repeatable, and more accurate for capital project participants. This framework and the toolsets are validated through the development of a practice conformance model, case studies on real project data, and application experiments. It is concluded that the proposed semi-automated framework for measuring practice conformance for capital projects is practical to implement in the near term. These results provide a basis on which capital project participants can implement efficacious practice conformance measurement to support capital project performance improvement programs.