Page 77 - The Six Sigma Project Planner
P. 77

Estimating Project Duration Statistically
                    While useful, the above analysis is missing some important information, namely
                    probabilities associated with each schedule. We know for example that the best- and
                    worst-case scenarios are combinations of improbable events and are therefore extremely
                    unlikely. These estimates provide useful bounds for our schedule estimates and are
                    helpful in determining whether or not the deadline is even realistic, but it would be
                    even better if we established a statistical distribution of schedule completion dates. We
                    will do so now.
                    Consider the schedule information in the table below, which is an excerpt from the table
                    used in the previous example. You may wish to review the diagram presented earlier to
                    confirm that critical path is A-B-C-D-G and the noncritical path is A-B-E-F-G. The
                    activities in gray cells are not on the critical path.

                                                            Activity Duration
                                           Activity
                                                     Mean      Variance     Sigma
                                             A        2.00       0.11        0.33
                                             B        5.33       1.00        1.00

                                             C        8.00       4.00        2.00
                                             D        6.33       4.00        2.00
                                             E        3.33       1.00        1.00
                                             F        4.50       0.69        0.83

                                             G        6.17       2.25        1.50

                    From these data it is possible to compute the mean, variance, and standard deviation for
                    the critical and noncritical paths. The path mean is the sum of the activity means (we
                    are using the weighted averages here), the path variance is the sum of the variances of
                    the activities on the path, and the path standard deviation is the square root of the path
                    variance. For these data we get the following statistical estimates:

                                            Path      Mean      Variance    Sigma

                                        ABCDG         27.83      11.36     3.370625
                                        ABEDF         21.33        5.06    2.248456

                    Statistically, due to the central limit theorem, the sum of five or more distributions will
                    usually be approximately normally distributed. Thus, for the critical path (and for the
                    project), the time to our scheduled completion date can be considered to be approxi-
                    mately normally distributed, with a mean time to completion of 27.83 working days and
                    a standard deviation of 3.4 working days. Reviewing the calendar for this project,
                    scheduled to begin on Monday, March 31, 2003, we see that (assuming resource



                                                             60
   72   73   74   75   76   77   78   79   80   81   82