Forecasting Construction Project Performance with Momentum Using Singularity Functions in LPS

Abstract

Construction planning and control are crucial for project success. The last planner system (LPS) presents a proactive approach to plan and control production, designed to increase reliability and enhance the make-ready process. Several metrics guide the planning process at the macro level (master and phase scheduling) and the micro level (lookahead and weekly work planning). However, LPS still lacks a mathematical model that can systematically and continuously analyze such metrics, especially to forecast project performance. Moreover, there are no studies on the effect of the fluctuations of lookahead-planning LPS metrics on the metrics at the weekly work plan level. This research, therefore, proposed a new mathematical model using singularity functions, which are types of range-based expressions that track the different paths that each task can follow, from lookahead planning to weekly work planning, and evaluate LPS metrics. To assess project performance, the concept of momentum was introduced as the rate of change in metrics from week to week. Momentum was applied to the Tasks Made Ready (TMR) metric to predict the Percent Plan Complete (PPC). Through machine learning models, results show that momentum can predict PPC with over 93% correlation between actual and predicted PPC values. Data from actual construction execution in the United States were used to validate the proposed model. The contribution of this research lies in (1) conceiving a mathematical model method for project control; and (2) introducing the concept of momentum, which takes the rate of change of any metric into account, incorporated into the LPS for more reliable planning. The methodology proposed in this study can help industry better plan its projects and leverage the concept of momentum to better predict PPC, which is essential for every planning and control process in construction projects.