Making each workhour count: Improving the prediction of construction durations and resource allocations

Abstract

Construction duration is an important performance aspect of building projects because it determines the time-to-market for a building, i.e., it stands between a client’s final decision to construct a building and obtaining the benefits from the designed building. This paper shows how intelligent computing (including semantic modeling, simulation, genetic algorithms, and machine learning) improves the ability of construction professionals to predict the construction schedule duration and direct cost of building projects at the beginning of construction and during construction. Such predictions are important because they inform the schedule commitments made and the allocation of resources that practitioners believe will let them meet the commitments. Hence, the prediction methods must capture the most important phenomena that are likely to impact the duration of activities and of construction. The two applications discussed – Tri-Constraint Method (TCM) and Activity-Flow Model (AFM) – incorporate key phenomena observed in practice, such as handling of workspace constraints and flows required for the execution of activities, that are not part of the currently prevalent concepts and tools. TCM and AFM significantly improve the ability of construction professionals to make more reliable predictions of construction duration.