Glossary · approach
Program Evaluation and Review Technique
Probabilistic extension of CPM developed in 1958 for the US Navy Polaris program: each activity duration is a random variable from optimistic, most-likely and pessimistic estimates, yielding an expected project-completion distribution.
PERTProgram Evaluation Review TechniqueThree-Time Estimate Technique
The Program Evaluation and Review Technique (PERT) is the probabilistic extension of CPM, developed by Malcolm, Roseboom, Clark and Fazar (1959) at the US Navy Special Projects Office for the Polaris fleet ballistic missile program. Each activity duration is treated as a random variable; for each activity an optimistic estimate a, a most-likely estimate m and a pessimistic estimate b are collected. Using a beta-distribution approximation the expected duration is computed as (a + 4m + b) / 6 and the variance as ((b - a) / 6) squared. Forward and backward passes on the activity graph are then performed using expected durations, and the variance of the critical path is the sum of variances of activities on it. By the central limit theorem the project completion time is approximately normally distributed, which lets the planner compute the probability of finishing by any target date. For practical projects with many parallel near-critical paths the analytical formula understates true variance and Monte Carlo simulation is preferred. Reference: Malcolm, Roseboom, Clark and Fazar 1959; Hillier and Lieberman 2014.
Örnek
A large industrial-plant project lists 120 activities, each with a three-time estimate; the PERT analysis returns an expected duration of 540 days with a standard deviation of 28 days on the critical path, giving the planner a 78% probability of finishing within the contractual 580-day window — and a 22% schedule-risk number to manage.