Different statistical process control (SPC) approaches were proposed over the years for project management using earned value management/earned schedule. A detailed examination of these approaches has led us to express a need for a unified framework in which to test and compare them. The main drivers for this need were the lack of a formal definition for a state of control, the unavailability of a benchmark dataset, the absence of measures to quantify the SPC performance and the lack of consensus on how to overcome and test the normality assumption. In this paper, we present such a framework that combines a classification from empirical data, a known project dataset, a sound simulation model and two quantitative measures for project control efficiency. Four SPC approaches from prior literature have been implemented and an exhaustive experiment was set up to compare and to discuss their value for the project management practice.

Simulation has played an important role in project-management studies of the last decades, but in order for them to produce practical results, a realistic distribution model for activity durations is indispensable. The construction industry often has needed historical records of project executions, to serve as inputs to the distribution models, but a clearly outlined calibration procedure is not always readily available, nor are their results readily interpretable. This study seeks to illustrate how data from the construction industry can be used to derive realistic input distributions. Therefore, the Parkinson simulation model with a lognormal core is applied to a large empirical dataset from the literature and the results are described. From a discussion of these results, an empirical classification of project executions is presented. Three possible uses are presented for the calibration procedure and the classification in project management simulation studies. These were validated using a case study of a construction company.

Recent literature on project management has emphasised the effort which is spent by the management team during the project control process. Based on this effort, a functional distinction can be made between a top down and a bottom up project control approach. A top down control approach refers to the use of a project control system that generates project based performance metrics to give a general overview of the project performance. Actions are triggered based on these general performance metrics, which need further investigation to detect problems at the activity level. A bottom up project control system refers to a system in which detailed activity information needs to be available constantly during the project control process, which requires more effort. In this research, we propose two new project control approaches, which combines elements of both top down and bottom up control. To this end, we integrate the earned value management/earned schedule (EVM/ES) method with multiple control points inspired by critical chain/buffer management (CC/BM). We show how the EVM/ES control approach is complementary with the concept of buffers and how they can improve the project control process when cleverly combined. These combined top down approaches overcome some of the drawbacks of traditional EVM/ES mentioned in the literature, while minimally increasing the effort spent by the project manager. A large computational experiment is set up to test the approach against other control procedures within a broad range of simulated dynamic project progress situations.

This paper explores the use of multivariate regression methods for project schedule control within a statistical project control framework. These multivariate regression methods monitor the activity level performance of an ongoing project from the earned value management/earned schedule (EVM/ES) observations that are made at a high level of the work breakdown structure (WBS). These estimates can be used to calculate the longest path in the project and to produce warning signals for project schedule control. The effort that is spent by the project manager is thereby reduced, since a drill-down of the WBS is no longer required for every review period. An extensive computational experiment was set up to test and compare four distinct multivariate regression methods on a database of project networks. The kernel principal component regression method, when used with a radial base function kernel, was found to outperform the other presented regression methods.

Project control has been a research topic since decades that attracts both academics and practitioners. Project control systems indicate the direction of change in preliminary planning variables compared with actual performance. In case their current project performance deviates from the planned performance, a warning is indicated by the system in order to take corrective actions. Earned value management/earned schedule (EVM/ES) systems have played a central role in project control, and provide straightforward key performance metrics that measure the deviations between planned and actual performance in terms of time and cost. In this paper, a new statistical project control procedure sets tolerance limits to improve the discriminative power between progress situations that are either statistically likely or less likely to occur under the project baseline schedule. In this research, the tolerance limits are derived from subjective estimates for the activity durations of the project. Using the existing and commonly known EVM/ES metrics, the resulting project control charts will have an improved ability to trigger actions when variation in a project׳s progress exceeds certain predefined thresholds A computational experiment has been set up to test the ability of these statistical project control charts to discriminate between variations that are either acceptable or unacceptable in the duration of the individual activities. The computational experiments compare the use of statistical tolerance limits with traditional earned value management thresholds and validate their power to report warning signals when projects tend to deviate significantly from the baseline schedule.