State of the Art of Correlation-Based Models of Project Scheduling Networks

In traditional network critical path method/program evaluation and review technique models for project scheduling, it is assumed that the durations or costs of individual activities are independently distributed. In real situations, however, there may exist correlations between parallel or serial activities; they may be positive or negative, linear, or nonlinear. The purpose of this paper is to examine how the correlations influence uncertain networks of project scheduling, and to summarize emerging research in this area. We review previous approaches modeling correlations or interdependence between activities by classifying them into three broad methodological groups: analytical models, simulation methods, and expert systems. These approaches are compared in terms of methodological approach, basic assumptions, inputs, advantages, limitations, and applications. Some new methods and areas based on complex system theory such as agent-based modeling and simulation (ABMS), system dynamics (SD) modeling, and integrated ABMS-SD modeling are noted as an agenda for analyzing or simulating correlations between activities in project scheduling.