A Practical and Accurate Alternative to PERT
Many real world projects can be represented by stochastic activity network (SAN) models, where the duration of some or all of the project activities are, at best, known in probabilistic sense. These models are known in the literature as PERT networks and they are analyzed using the PERT procedure. It has been known for many years that the Project Evaluation and Review Technique (PERT) provides inaccurate information about the project completion time. Quite often this inaccuracy is large enough to render such estimates as not helpful. As a result of this inaccuracy, many improvements since the introduction of PERT in 1959 have been developed. However, in spite of this inaccuracy and the many improvements, PERT procedure continues to be taught and presented in most text books on Project Management. This is due, perhaps, to its simplicity, and ease of its application. In this paper a new alternative is developed that addresses the issues of accuracy and practicality simultaneously in analyzing SAN models. The new procedure is based on a more accurate representation of the distribution function of the project completion time. We first show that the project completion time can in certain instances be accurately represented by a normal distribution, but in many other instances it can not. In these other instances we show that the project completion time can be more accurately represented by an extreme value distribution. Hence, SAN is first characterized as to when we can use the normal distribution and when we can use extreme value distribution. In the first case, PERT estimates are accurate and will be used; however, in the second case a new procedure is developed that is easy to use, and results in more accurate estimates of the project completion time and its statistics. In this paper examples are also provided that illustrate the above characterization of SANs and the accuracy and practicality of the new procedure.
KeywordsStochastic activity networks extreme value distribution normal distribution project completion time estimation
Unable to display preview. Download preview PDF.
- Adlakha, V. G., and Kulkarni, V. G. (1989). A Classified Bibliography of Research on Stochastic PERT Networks: 1966–1987, INFOR, 27(3):272–296.Google Scholar
- Burt Jr., J. M. and Garman, M. B. (1971). Conditional Monte Carlo: a Simulation Technique for stochastic network analysis, Management Science, 18:207–217.Google Scholar
- Clark, C. E. (1961). The Greatest of a Finite Set of Random Variables, Operations Research, 9:146–162.Google Scholar
- Clingen, C. T. (1964). A Modification M Fulkerson’s PERT Algorithm. Letter to the Editor, Operations Research, 12:629–632.Google Scholar
- Cramer, H. (1946). Mathematical Models of Statistics, Princeton University Press, Princeton, NJ.Google Scholar
- Demeulemeester, E., Herroelen, W. (2002). Project Scheduling: A Research Handbook, Kluwer Academic Publishing.Google Scholar
- Dodin, B. M. (1985). Bounding the Project Completion Time Distribution in PERT Networks, Operations Research, 24:862–882.Google Scholar
- Donaldson, W. A. (1965). Estimation of the Mean and Variance of a PERT Activity Time, Operations Research, 13:382–385.Google Scholar
- Dubois, D. and Prade, H. (1987). Fuzzy Numbers: An Overview, in: Analysis of Fuzzy Information, Bezdek, J.C., ed, CRC Press, Boca Raton, pp. 3–39.Google Scholar
- Fulkerson, D.R. (1962). Expected critical path lengths in PERT networks, Operations Research, 10:808–817.Google Scholar
- Glover, F., Klingman, D., and Phillips, N. V. (1992). Network Models in Optimization and their Applications in Practice, John Wiley & Sons, New York.Google Scholar
- Van Slyke, R M. (1963). Monte Carlo methods and the PERT problem, Operations Research, 11:839–860.Google Scholar