Guild of Project Controls: Compendium | Roles | Assessment | Certifications | Membership

Tips on using this forum..

(1) Explain your problem, don't simply post "This isn't working". What were you doing when you faced the problem? What have you tried to resolve - did you look for a solution using "Search" ? Has it happened just once or several times?

(2) It's also good to get feedback when a solution is found, return to the original post to explain how it was resolved so that more people can also use the results.

critical path.

18 replies [Last post]
Thu, 2003-11-06 02:26
Dayanidhi Dhandapany
User offline. Last seen 3 years 3 weeks ago. Offline
Joined: 18 Mar 2003
Posts: 470
Groups: None
Which relationship like FS or FF (with positive lag) is the best practical way to use in order to form a critical path? in my opinion all activities must have finish to start relationship if we have enough detailed break down. During the absence of such details(delay in awarding sub-contract works)what could be done? i leave this question to our fellow planners.....................

What is the easiet way to group multiple critical paths in P3. Other than Navigating through driving relationships and assign some custom codings corresponding to each critical paths..............
Top

Replies

Mon, 2005-01-03 14:59
#1
Bernard Ertl
User offline. Last seen 9 years 18 weeks ago. Offline
Joined: 20 Nov 2002
Posts: 757
Groups: Petro-Chem / Oil & Gas, TURNAROUND & SHUTDOWN (T/A & S/D)
Unfortunately, it is still deficient and even incorrect. ... To see a quick difference, compare your hand calculations to a similar schedule in P3....We discussed this in the CPM thread. As I explained in my last post there (and quoting from your paper), the algorithms produce the same schedules (calculated dates). The difference is only in where the offset is handled.

Getting back to your original statement. Your ADM examples do not match PDM. ADM cannot note the start of any activity, only the finish. Your definition depends upon the finish of the predecessor activity to logically start your Start-to-Start successor. It does not take into account that the successor ADM activity may not have actually started. It does not consider that there may be multiple predecessors to the original successor activity.Please define what you mean by "note the start of an activity". We discussed the ADM/PDM start to start (SS) constructs in the ADM Software thread and I’m still convinced that your argument is not congruent with the definition of the ADM calculations (see below for further clarification).

Finally, status is not an “implementation consideration.” The data (or status) date must be taken into account if you want the schedule to model reality. The article does not mention using calendars (especially multiple calendars.) It does not address interruptible relationships (which is similar but different from the activity splitting mentioned.) Without status, you cannot begin to address the complexities of computing the CPM with out-of-sequence progress.As you mentioned, resource leveling, disjoint calendars, etc. are not considered in the definition of the ADM & PDM calculations (in the referenced book or any text that I am aware of). They are extensions of the defined calculations and there is no standard for handling/calculating them. Their implementation in any given software application will be set according to the software engineer’s algorithm(s). As such, they are not really germane to a discussion on scheduling that is limited to ADM & PDM definitions IMO.

Modern CPM is much more complicated that even the college professors know. Only a handful of engineers understand the complexities of a complete CPM system (and they mostly are working under confidentiality agreements!)I have worked on the CPM engine in our ATC system for several years. I think I’m pretty well aware of the complexities of programming a complete CPM system.

I’m guessing that Mr. Liberzon is also in that group. Perhaps he will weigh in with his thoughts.

It is time that more people became involved discussing these issues, as you are doing here. Keep up the good work!Cheers Ron. I agree wholeheartedly and appreciate your part in the discussions.

Bernard Ertl
ATC Professional Shutdown / Turnaround Management System
Top
Mon, 2005-01-03 13:51
#2
Ronald Winter
User offline. Last seen 3 years 5 weeks ago. Offline
Joined: 4 Jan 2003
Posts: 928
Groups: None
The article that you quote is an excellent one – perhaps the best one available. Everyone should have this paper and refer to it. Unfortunately, it is still deficient and even incorrect. It uses the CPM calculation method that I have come to term, “Aggie Method” and described in a published article in the Cost Engineering journal of AACEi. I have posed a version of this article (“How to Befuddle a College Professor without Really Trying”) on my website at http://www.ronwinterconsulting.com/published.htm. I believe that Microsoft Project uses this method.

Primavera (in all of its incarnations) does not use this method. They use the method called the “AACEi Method.” To see a quick difference, compare your hand calculations to a similar schedule in P3. Be sure to look at the work days. You will see that they are different. No matter how simple your example, if you look, you will see differences in workdays for every activity.

Getting back to your original statement. Your ADM examples do not match PDM. ADM cannot note the start of any activity, only the finish. Your definition depends upon the finish of the predecessor activity to logically start your Start-to-Start successor. It does not take into account that the successor ADM activity may not have actually started. It does not consider that there may be multiple predecessors to the original successor activity.

Finally, status is not an “implementation consideration.” The data (or status) date must be taken into account if you want the schedule to model reality. The article does not mention using calendars (especially multiple calendars.) It does not address interruptible relationships (which is similar but different from the activity splitting mentioned.) Without status, you cannot begin to address the complexities of computing the CPM with out-of-sequence progress.

Modern CPM is much more complicated that even the college professors know. Only a handful of engineers understand the complexities of a complete CPM system (and they mostly are working under confidentiality agreements!) It is time that more people became involved discussing these issues, as you are doing here. Keep up the good work!
Top
Mon, 2005-01-03 12:10
#3
Ronald Winter
User offline. Last seen 3 years 5 weeks ago. Offline
Joined: 4 Jan 2003
Posts: 928
Groups: None
Becase, Bernard you are only looking at an un-statused baseline schedule. PDM operates differently from ADM when status is applied, especially when the actity is in-progress.
Top
Mon, 2005-01-03 10:06
#4
Bernard Ertl
User offline. Last seen 9 years 18 weeks ago. Offline
Joined: 20 Nov 2002
Posts: 757
Groups: Petro-Chem / Oil & Gas, TURNAROUND & SHUTDOWN (T/A & S/D)
Sorry to drudge up an old thread, but I wanted to clarify my original post to this thread. I realized that there is in fact a logical equivalent in ADM for PDM FF & SF ties.
  • FF (finish to finish) - Can be expressed as a dummy from the predecessor’s end node to the successor’s start node with a duration equal to the successor’s duration * -1.
  • SF (start to finish) - Can be expressed as a dummy from the predecessor’s start node to the successor’s start node with a duration equal to the successor’s duration * -1.
Unless a software system is designed to assist the scheduler by keeping the durations correct through updates automatically, this would be a nightmare to maintain in ADM.

I hereby acknowledge my previous mistake and concede that since a PDM schedule can, in fact, be modeled with ADM, all PDM schedules should be compatible with the critical path method. Why didn’t someone correct me earlier?

Bernard Ertl
eTaskMaker Project Planning Software
Top
Mon, 2005-01-03 12:29
#5
Bernard Ertl
User offline. Last seen 9 years 18 weeks ago. Offline
Joined: 20 Nov 2002
Posts: 757
Groups: Petro-Chem / Oil & Gas, TURNAROUND & SHUTDOWN (T/A & S/D)
Ron, I based my conclusions above based upon the descriptions of the PDM and ADM calculations described here. The descriptions of the calculations for ES, EF, LS, LF are independent of baselines, progress handling, etc. The differences you describe are dependent upon application specific implementations of the calculations IMO.

Bernard Ertl
eTaskMaker Project Planning Software
Top
Tue, 2004-03-16 18:26
#6
Vladimir Liberzon
User offline. Last seen 16 weeks 3 days ago. Offline
Joined: 5 Feb 2001
Posts: 2748
Groups: Accreditation Editors, Board of Advisors, Civil Engineering, GPCCAR Editors, Petro-Chem / Oil & Gas, Regional Representatives, Spider Project Team, TURNAROUND & SHUTDOWN (T/A & S/D)
Luca,
please write me a letter to spider@mail.cnt.ru.
I will send you back the examples.
Vladimir
Top
Tue, 2004-03-16 17:37
#7
Vladimir Liberzon
User offline. Last seen 16 weeks 3 days ago. Offline
Joined: 5 Feb 2001
Posts: 2748
Groups: Accreditation Editors, Board of Advisors, Civil Engineering, GPCCAR Editors, Petro-Chem / Oil & Gas, Regional Representatives, Spider Project Team, TURNAROUND & SHUTDOWN (T/A & S/D)
Luca, in Spider Project (like in P3) you can choose which resources to level. For others the package will calculate requirements.
Spider Project calculates expenses and revenues, you can simulate penalties and thus you are able to estimate what decision is the best - to use more resources or to pay penalties.
Next time we will include in our Demo sample projects for which Spider Project schedule will be shorter than produced by P3 and other packages, Spider Project optimizes resource constrained schedules. The example you tried is too simple and other packages produce the same results.
Spider Project calculates real feasible floats taking into account resource constraints not only on forward (like other packages) but also on backward pass. These floats can be used without damaging project duration.
Resource driven duration in P3 is not similar to productivity usage. Adding resources you shorten activity duration but you can not define different productivities for different resources, you cannot define resource productivities for typical activities (norms) and create references.
There are other useful and simple features - possibility to set any number of cost components (not the same as cost accounts), to use different currencies for different cost components, to create any number of cost centers (sum of expenses by defined set of components), to level by financial and supply constraints (revenue and supply simulation).
Another thing I suggest to try is skill scheduling (we call it resource assignment pools). And risk simulation too - this thing is very useful.
I will send you a couple of sample projects illustrating these features. Sample project in Demo illustrates our approach to risk simulation and management by success probability trends.
Regards,
Vladimir
Top
Tue, 2004-03-16 02:34
#8
Luca Basile
User offline. Last seen 9 years 10 weeks ago. Offline
Joined: 10 Jul 2003
Posts: 288
Groups: TILOS
I give a look at all the material.
But if I well understood the package is making a resoruce levelling taking in consideration only the max resoruce constrain (availability in time). And this mean 99% of the time delay the project.
Sometimes is mandatory to complete the project in time, as the penalties can be bigger than hire more people or plant for some periods.
What I saw P3 can also making levelling schedule taking like major contrain the project completion and trying to accomodate the resoruce to produce less overload possible.

I try to do the example in P3, and the results are matching, with exception for the total float estimation.
How are You estimating it?

Is great that the duration is based on the productivity.
But something similar is possible to do also in P3, when I have a resoruce driven resoruce.

But what I really appreciate is the opportunity to have different Wbs set up, as my need can be different from the Client or from other departments. As well I can want to have a Wbs function or product oriented for different report pourpose.

Top
Mon, 2004-03-08 04:04
#9
Vladimir Liberzon
User offline. Last seen 16 weeks 3 days ago. Offline
Joined: 5 Feb 2001
Posts: 2748
Groups: Accreditation Editors, Board of Advisors, Civil Engineering, GPCCAR Editors, Petro-Chem / Oil & Gas, Regional Representatives, Spider Project Team, TURNAROUND & SHUTDOWN (T/A & S/D)
Dear Luca,
Spider Project Demo can be downloaded from http://www.spiderproject.ru/spider_e.shtml<
And I advise to read presentations and papers that are presented at http://www.spiderproject.ru/publ_r.shtml at the bottom of the page (in English). In these presentations the concept of Resource Critical Path (RCP) is discussed and some examples are presented.
RCP is calculated by levelling resources not only on forward but also on backward pass. As the result you will receive activity floats that are calculated taking into consideration resource constraints. These floats can be used! Activity total floats calculated by American packages are not feasible because these packages do not level resources on a backward pass. RCP consists of activities with these resource constrained floats less than specified amount (usually zero, defined by user).
There are many other useful and unique features in Spider Project desribed in some presentations at this site.
Please pay attention that Spider Project optimizes resource constrained schedules and thus they are usually shorter than produced by other software due to better resource utilization. Besides you can level not only by resource constraints but also by financial and supply constraints.
Please don’t hesitate to ask any questions you may have.
Best Regards.
Top
Mon, 2004-03-08 03:23
#10
Luca Basile
User offline. Last seen 9 years 10 weeks ago. Offline
Joined: 10 Jul 2003
Posts: 288
Groups: TILOS
Dear Vladimir I find the spider project on internet.
I give a look at the resource CP.

Is similar a resource leveling, when I have resource constrain in quantity and/or time frame.
Just You consider like priority the resoruce, but this is always. I have always to identify the key resoruce and then
level on it or them.

As well I find the power point slide about your new software. There is any demo available to be downloaded?


Top
Sat, 2004-03-06 02:43
#11
Vladimir Liberzon
User offline. Last seen 16 weeks 3 days ago. Offline
Joined: 5 Feb 2001
Posts: 2748
Groups: Accreditation Editors, Board of Advisors, Civil Engineering, GPCCAR Editors, Petro-Chem / Oil & Gas, Regional Representatives, Spider Project Team, TURNAROUND & SHUTDOWN (T/A & S/D)
If resources, financing or supplies are constrained then traditional critical path does not make sense. We use the notion of Resource Critical Path consisting of activities that have zero resource constrained total float. This float is calculated when not only forward but also backward pass takes into consideration resource constraints. RCP is the same as Critical Chain if to add financing and supplies constraints to Critical Chain definition. This technique is used in Russia since 1993, long before Critical Chain approach was suggested. RCP can consist of activities that are not linked with each other.
Top
Thu, 2004-03-04 16:22
#12
David Watters
User offline. Last seen 19 years 4 days ago. Offline
Joined: 6 Jan 2003
Posts: 51
Groups: None
Luca

To determine Critical Path on Primaver, you must specify all activites with total float less than 1. Specifying activities with total float less than 0 will only bring up those activities that are super critical, ie are busting the end date

Cheers
Dave W
Top
Thu, 2004-03-04 05:44
#13
Darrell ODea
User offline. Last seen 18 years 10 weeks ago. Offline
Joined: 27 Feb 2002
Posts: 174
Groups: None
critical path.

You Wrote,
Asking a question which you have answered.
FS, SS & FF Relationships are the most practical way to find a CP.
In identifying CP & groups of CP, if you use a filer which identifies total float as less than or equal to 0.
--------------------------------------------------------------------------------
Which relationship like FS or FF (with positive lag) is the best practical way to use in order to form a critical path? in my opinion all activities must have finish to start relationship if we have enough detailed break down. During the absence of such details(delay in awarding sub-contract works)what could be done? i leave this question to our fellow planners.....................

What is the easiet way to group multiple critical paths in P3. Other than Navigating through driving relationships and assign some custom codings corresponding to each critical paths..............
Top
Tue, 2004-02-24 02:33
#14
Luca Basile
User offline. Last seen 9 years 10 weeks ago. Offline
Joined: 10 Jul 2003
Posts: 288
Groups: TILOS
Can You give me same idea how to do with P3, and where to find some material to read on it?

Thanks
Top
Sat, 2004-02-21 12:44
#15
Mehdi Rashidi Ala...
User offline. Last seen 5 years 37 weeks ago. Offline
Joined: 8 Jul 2003
Posts: 203
Groups: Building Information Modelling, Petro-Chem / Oil & Gas, Planning, Claims & Contracting
Dear my friends,
I think this item is old technique.
Are there any technique ( such as crtitical group method or ... ) in the Pm?
Top
Thu, 2004-02-19 02:47
#16
Luca Basile
User offline. Last seen 9 years 10 weeks ago. Offline
Joined: 10 Jul 2003
Posts: 288
Groups: TILOS
I do not think is useful to coding the critical path/s as it/they can change from daily to monthly bases.
It depend from the kind of programme.
Take in consideration a shut-down/maintenanece programme, that run in few days/weeks and is monitored at least once day. So every day you change the coding,to identify the critical path?

The critical path you mentioned have TF zero or are "iper-critical" one with TF<0?
If the TF is less than 0, each iper-critical path will have it representative TF, so You can filter or organize on it.
Pay attention if You are using a multicalendar programme, sometimes the critical activities seat on the longest path!

You can use also the RA applications. If You identify an activity of the critical path, there is a tools that is telling you all its predecessor and successor.

Top
Sun, 2004-01-11 12:02
#17
Bernard Ertl
User offline. Last seen 9 years 18 weeks ago. Offline
Joined: 20 Nov 2002
Posts: 757
Groups: Petro-Chem / Oil & Gas, TURNAROUND & SHUTDOWN (T/A & S/D)
For those who are interested, a side discussion on the whether or not a PDM SS relationship can be modeled equivalently in ADM is ongoing in this thread:

ADM / PDM SS relationships thread

Bernard Ertl
InterPlan Systems Inc. - Project Management Software, Project Planning Software
Top
Thu, 2003-11-06 10:25
#18
Bernard Ertl
User offline. Last seen 9 years 18 weeks ago. Offline
Joined: 20 Nov 2002
Posts: 757
Groups: Petro-Chem / Oil & Gas, TURNAROUND & SHUTDOWN (T/A & S/D)
Critical Path Method - A technique used to predict project duration by analyzing which sequence of activities has the least amount of scheduling flexibility. Early dates are figured by a forward pass using a specific start date and late dates are figured by using a backward pass starting from a completion date.

There are other definitions, but this one is appears to be the most common.



"Classic CPM theory would calculate the early start date from the project’s notice to proceed or data date. Many schedulers consider the classic theory the proper calculation method for the determination of early start and early finish, The issue revolves around the selection of which float values to use in precedence schedules and proper relationship rules. Start float, activity float, and finish float can vary dramatically, with complex relationships. However, under classic scheduling theory, there can be one early start float, multiple activity floats for different portions of the activity, and one finish float." - excerpted from Construction Delay Claims by Bramble and Callahan

IOW, calculate early starts, early finishes and late finishes in one direction (each).



To answer your question, it may be instructive to look at the differences between ADM (arrow diagramming method) and PDM (precedence diagramming method). ADM and PDM are not logical equivalents. PDM can fully describe ADM networks, but the complement is not true. PDM contains constructs that can not be adequately described by ADM (and thus, are of debatable application in a strict critical path method schedule, IMHO).
  • FS (finish to start) - logical equivalent in ADM (predecessors end node = successors start node)
  • SS (start to start) - logical equivalent in ADM (dummy task connects predecessors start node to successors start node)
  • FF (finish to finish) - no exact equivalent in ADM. Can be expressed in ADM with a dummy task connecting the predecessors end node to the successors end node, however, this does not produce the same result as a FF tie in PDM because ADM logic networks drive relationships in one direction. The resulting ADM schedule will not force a Gantt chart to display the tasks as completing at the same time, just that the successors succeeding tasks are dependent upon both.
  • SF (start to finish) - no logical equivalent in ADM. ADM logic networks drive relationships in one direction.
  • LAGS (FS & SS) - logical equivalent in ADM using dummy tasks with durations equal to the lag span.
  • LAGS (SF & FF) - no logical equivalent.

The critical path method was created with ADM. Straying from logical equivalents of that construct may lead to schedules that do not meet the classic definition of a critical path method schedule.

Bernard Ertl
InterPlan Systems Inc. - Project Management Software, Project Planning Software
Top