Pert Analysis

[John Volpe]

I've used the PERT analysis tool to calculate expected,
optimistic and pessimistic durations and it worked fine.

Can anyone tell me how to display the associated costs
for each of these three scenarios? I tried adding a
column and picked the 'Cost' field, but it only displays
the cost from the Gantt chart (i.e., the cost is the same
for all three scenarios).

Thanks!

 

[JackD]

I don't believe it calculates cost.
depending on how your resources are assigned, it could be as simple as
copying the pessimistic duration to the duration column and recalculating,
then copying the cost into one of the spare cost fields. THen doing the same
for optimistic. However, if your resource assignments are
contoured/complicated and set to something like fixed work, then it may not
work as well.

I suggest making a small test file (representative of your project but with
a small number - <10 - of activities) and experimenting to see if it returns
the results you are expecting. Using a few activities let's you calculate by
hand what you expect so you can confirm that the project results are
correct.

-Jack

 

[Mark Durrenberger]

Careful John,

Don't add up all the pessimistic values or all the optimistic values. The
addition of these will result in mathematical nonsense. Why?

The PERT technique uses a range of durations for each activity. The
optimistic duration the most likely duration, and the pessimistic duration.
Depending on who you ask (or what reference you read) when you estimate the
optimistic duration and the pessimistic duration, you are supposed to
estimate a duration with as little as 1 in 100 or 5 in 100 likelihood of
occurring (on the optimistic side) or 95 or 99 in 100 chance of occuring on
the pessimistic side.

For simplicity, let's assume your optimistic duration estimates are all of
the "5%" variety (5 in 100 chance of occurring) and you decide to add them
up. Essentially what you've done is add up a bunch of 5% events which
results in a really really really low likelihood schedule


So why bother with PERT? The important results from PERT come from the
range estimates. Using the optimistic, most likely and pessimistic
durations, you calculate a Mean duration (Mean = average)
with the equation (op + 4ML + pess)/6 (PERT assumes a beta distribution, an
approximation of the mean for a beta distribution is the equation above) if
you add up these Mean durations (for the critical path), you will get a
project duration mean with 50% likelihood of success.

Cost works similarly, however you must add up range cost estimates for each
task (with duration, you only add up range duration estimates for the
critical path activities).

http://www.oakinc.com/pdf/estimates.pdf - this article might help a bit. It
explains a technique similar to PERT (using a triangle distribution for
duration instead of the beta distribution assumed by the PERT technique)

Happy to answer any other questions,
Regards,
Mark



--
_________________________________________________________
Mark Durrenberger, PMP
Principal, Oak Associates, Inc, www.oakinc.com
"Advancing the Theory and Practice of Project Management"
________________________________________________________

The nicest thing about NOT planning is that failure
comes as a complete surprise and is not preceded by
a period of worry and depression.

- Sir John Harvey-Jones