Progress monitoring of construction projects using pattern recognition techniques

Abstract

Project monitoring involves collecting the actual‐progress data, and comparing them against the relevant planned‐progress data to evaluate the overall project progress at specified cut‐off dates. Inevitable issues including variations in reporting skills as well as the willingness to record accurate data impact on the quality of the collected data. Comparison against multiple possible benchmarks (one‐to‐many) rather than a single benchmark (one‐to‐one) offers the potential to alleviate the negative impact of low‐quality data on the progress evaluation. Special patterns, which can be readily manipulated within computer programs, are devised to encode the planned and actual progress at the cut‐off dates. Basically, pattern recognition techniques are utilized to classify the multiple patterns representing the planned progress at a given cut‐off date and the classification is used to evaluate the pattern representing the actual progress at the same date. The pattern recognition techniques generalize a virtual benchmark to represent the planned progress based on multiple patterns generated at a given cut‐off date and representing possible benchmarks. In addition to the alleviation of the negative impact of low‐quality data on the progress evaluation, the generalization feature potentially encourages a long‐run attitude in site personnel to report high‐quality data. Finally, the pattern recognition concept and technique proved their robustness to monitor and evaluate the overall progress of the projects based on the technique of critical path method.

Keywords:

• Progress monitoring
• project control
• pattern recognition
• artificial neural network
• statistical models