Critical Path

[Vincent Isoz]

Hi,

Since MS Project exists the critical path calculation is incorrect as far as
i know (and it's not the only false point...).

If you have a task with a positive total slack and put him a constraint like
"must start on" she will become "critical" (red) but mathematically she's not!

I don't understand why this has not been corrected since this last 10 years
O_O

The computer scientist by Microsoft have to revised the Dijkra algorithm in
ponderated and connex graph theory to calculate correctly the critical path
as Primavera do.

When will this be corrected... in 2 years... or 10 years again? I lot of my
customers are unsatisfied by this behavior.

Thanks for your answer

 

[Jim Aksel]

I can't speak for Microsoft, I don't work for them ...
However, I suggest that marking a task as MSO (Must Start On) and some of
the other constraints does make a task critical....

The basics of critical path (0 slack tasks) are well understood. Consider
these ...
Task A -- 10 days
Successort to A is Task B which is also 10 days.

I have task "C" which is 2 days in duration and has a MSO on project day 3
with a successor of Tasis B. Basic CPM gives us slack of 5 days (end of C to
start of B). But since you have MSO on C, the logic we just used to
calculate slack of 5 days goes away. If you insist the slack is 5 days, I
ask then why is task C MSO? How many days can you delay "C" without
violating the MSO constraint? That answer would be 0. Tasks with 0 slack
are on the critical path.

I certainly see your point, but I think there are more subtle things going
on in the calculation.

HTHIYW
--
If this post was helpful, please consider rating it.

Jim Aksel, MVP

Check out my blog for more information:
http://www.msprojectblog.com

 

[Vincent Isoz]

Thanks

But because of that top project manager having high mathematics skills in
management know that ms project don't follow official worldwide and correct
algorithms and the problem remains anyway. The CPM is false and stay false
and also completely useless in this software!

Other problem too... the calculation of the Total Slack and Late Finish is
false most of time too...

 

[Steve House]

I respectfully submit that you (or the aformentioned "top project manager"
in your organization) are the one who has the mistaken idea of what is the
proper way to calculate what constitutes a critical task.

 

[Vincent Isoz]

Hi

The correct way is the Dijkra algorithm in
ponderated and connex graph theory in the art of the highest summation
values path (see documentation in "Graph Theory" books).

This is the manner we learn to calculate it in scientific management.

 

[Steve House]

The ANSI accepted definition of critical path as shown in the PMBOK 3rd
Edition, includes the exception statament that the critical path may
terminate prior to the end of the project when it encounters a scheduled
milestone that has a fixed date due to the application of a hard constraint
such as MFO or MFNLT. Remember that unlike defintions within the physical
sciences, terms such as "critical" are human-devised terms chosen for
conveneience in communication and are not defined as laws of nature. Graph
theory may tell you one thing - I'mn not a mathematician and can't comment
one way or the other on what you were taught - but the curently accepted
definition document for professional practices in Critical Path Methodology
project management says that milestones can and should affect the critical
path in precisely the way you are complaining that MS Project displays it.
MS Project is in conformity to ANSI standards in that regard - your
preferences are not.

 

[Vincent Isoz]

Ok thanks for your answer.

I won't make any comment about definition of the PMBOK (that is far away to
be a reference for complex projects). But if the ANSI accepted it i have
nothing to say.

Otherwise i have a question for the specialists you are. How do you manage
yourself tasks in sharepoint that are show as critical but that have a
physicial slack not equal to zero (i don't speak about the calculate slack of
MS Project but the "physical slack" like Primavera calculates)?

Thanks a lot for your answer because i have to answer to my customers how to
manage this...

 

[Jim Aksel]

Vincent, perhaps you can explain to your customer that different software
packages implement critical path calculations in different ways. It appears
PV uses the algorithm more literally, where as Microsoft Project is more in
line with how ANSI/PMI and the general industry are interpreting the matter.

By the way, here is an interesting read:
http://www.gridbus.org/papers/DCP-eScience2007.pdf

Also, please do not confuse high power mathmatics with the scheduling
techniques used by the commercially available software (from various
vendors). I am reminded of the mathmatician who lets a pile of papers burn
on his desk and does nothing. The engineer, who puts the fire out with the
glass of water on the mathmaticians desk asks why he let the papers burn.
The mathmatician replied, "The solution was intuitively obvious."

So, what the theory says must be implemented considering practical and
acceptable measures.
--
If this post was helpful, please consider rating it.

Jim Aksel, MVP

Check out my blog for more information:
http://www.msprojectblog.com

 

[Gary L. Chefetz]

Vincent:

What is your algorithm for calculating Physical Slack?