Optimal Project Portfolio Selection Considering Cascading Failure Among Projects

Abstract

A novel model is proposed for optimal project portfolio selection (PPS). Different from previous PPS methods that consider the interdependence, interaction, or synergy relationships among projects, this article considers the cascading failure process caused by these relationships. Specifically, we first construct a PP network (PPN) and determine the initial load and capacity of the projects (nodes). Afterward, we simulate the cascading failure process by attacking one or more projects in the PP and distribute the load of the failed projects to nonfailed neighboring projects. The cascading failure process ends until no project fails. Finally, we propose a new indicator to measure and compare each PPN's cascading failure results, based on which the optimal PP is selected. A numerical experiment is conducted to depict the PPS process. The results of this study provide guiding implications for project portfolio management decision-makers, helping them consider PPS problems from a new perspective and make more reasonable decisions.