A distributionally robust analysis of the program evaluation and review technique

Abstract Traditionally, stochastic project planning problems are modeled using the Program Evaluation and Review Technique (PERT). PERT is an attractive technique that commonly used in practice as it requires specification of only a few characteristics activities’ duration. Moreover, its computational burden extremely low. Over years, four main disadvantages have been voiced much research has devoted to analyzing them. The effect beta distribution corresponding variance assumes investigated numerous studies, through results for variety other distributions. In this paper, we propose more general method PERT’s sensitivity assumptions regarding distribution. particular, do not assume singular activity duration, but instead be partially specified by support, mean possibly absolute deviation. exact worst- best-case expected durations over set distributions can calculated from distributionally robust optimization on themselves. A numerical study instances PSPLIB shows assumption underlying limited. Furthermore, find added value knowing deviation also modest.