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.

Resource Linking - good/bad?

17 replies [Last post]
Mark Chapman
User offline. Last seen 43 weeks 1 day ago. Offline
Joined: 15 Oct 2018
Posts: 3
Groups: None

I'm kicking off the planning of a 200,000 hr shutdown due in 12 months. Previous shut downs here have used extensive resource logic between physically unrelated activities as a way of manually resources.

I'm not a fan of resource linking, I'd rather use the automatic resource leveling and let any unrelated activies sit on data date if they are available for work to start.

I can see there are pros and cons for both methods, udateing with resource links can be a nightmare with work all over the place and possibly/most likely incorrect critical paths poping up at every update.

However not linking activities can make it difficult to undersand the "bow wave" of work being pushed out each day.

I'd like to ask the Planning Plannet gurus for their experiences in this area.

We will be using Primaver on this project.

Replies

Trevor Rabey
User offline. Last seen 4 weeks 6 days ago. Offline
Joined: 29 Nov 2005
Posts: 528
Groups: None

Mark, I am just gong to stick to your original question. What was it again? Well, I looked for a sentence with a question mark at the end of it and there isn't one, so I am guessing it is "what do you think?"

If you link Task A as a FS0 predecessor of Task B, you are making an explicit statement that Task B cannot start earlier than Task A finishes. There is nothing in it to indicate whyTask B cannot start earlier than Task A finishes. If there was, then attention could focus on removing that reason so that the link could be removed. There might be more than one reason for the predecessor link.

I prefer to avoid the fairly common practice of linking tasks as predecessor/successor if the reason is only because of resource limitations (say there is just one guy to perform both tasks) and there is a preference (but not a strong explicit reason) that the guy does Task A first. I prefer to only link Task A and Task B if they are something like "can't put the roof on until the walls are built", where the reason is something reliable and inescapable, such as gravity. Otherwise, you constantly find that someone finds a way to do Task B first, and goes ahead and does it, or more resources are found and someone decides to go ahead and do Task B as soon as it can be done, without reference to Task A at all. That's a bad thing because the original planning is being ignored and has to be undone and re-done, but it is a good thing because tasks get done earlier than expected (what a pleasant surprise). My way, you end up with resource over-allocation and have to let the software handle the resource levelling. And why not? After all, why buy a dog and then do the barking? The other way, it gets so complicated very quickly that a human planner/scheduler is overwhelmed by the complexity, in the planning and especially in the tracking and updating.

Anoon,

I wrote about Spider Project in response to your statement "I guess we all know that common software(s) are only capable of linear calculations." I do not want to discuss Spider Project here because the topic of this discussion is different and Mark uses and will use P6. So I will answer to your question and if you will have other questions let;s discuss them in PP Spider Project forum.

An example of variable resource assignments in Spider Project:

You may set that the job may be done by from 2 to 6 workers. If 2 will become available the work can start, if others will become available when this activity is not finished they may join and accelerate activity execution.

An example for skill scheduling:

You have several excavators with different capacities. You may assign the skill Excavator and required quantity and the software will select those excavators that are available at activity start and are optimal in accordance with the selected criteria (do the job cheaper, do the job faster, etc.).

An example for continuous activity:

You will not start concreting at the end of the day and finish at the beginning of the next day. If an activity is continuous Spider Project will delay its start to be able to finish it the same day.

I can suggest a lot of examples of smart planning realized in Spider Project. The software shall be smarter than the people. It shall take into account all existing constraints and create the best possible schedule. In other case people will plan manually and use the software only as communication tool.

It is not necessary to know Spider Project for questions like if it is possible to model some real life situations like described above. If you will suggest something that Spider cannot model I will appreciate it because your example will help us to improve Spider Project capabilities. We publish new versions each week adding new options. We do not plan to stop.

Regards,

Vladimir

Anoon Iimos
User offline. Last seen 18 weeks 3 days ago. Offline
Joined: 22 Sep 2006
Posts: 1416

Vladimir,

I can't, as in as much as I want to (make a challenge), I've never used Spider Project. So I guess it's your call to prove the superiority of your software as compared to others. I mean no sales talk. Just an honest to goodness presentation perhaps and having not much of esoteric terms. I guess the audience will love it! I can't follow Rafael as I eventually get to sleep when I read most of his posts. 

As you said, Spider can handle variable resource assignments and or variations automatically (automatic resource(s) leveling).

What do you mean exactly, and how are you going to do that using Spider Project? 

Apologies to Mark, however I guess this may able to help in some way, in solving problems regarding resource(s) leveling. 

And who knows? P6 users might switch to Spider eventually! 

Anoon,

for your information:

Spider Project captures variations (we call it variable resource assignments) and shifts and plans projects in minutes and seconds. I don't want to discuss its features here but if you want to challenge Spider Project with something it does not model create a discussion in Spider Project forum.

Anoon Iimos
User offline. Last seen 18 weeks 3 days ago. Offline
Joined: 22 Sep 2006
Posts: 1416
Again (simply guessing): When software calculates automatically, I believe it only distributes resource units/quantities equally against the Planning Unit - following logic links. It cannot capture variations and shifts. So for me, you need to structure and granulate your activities and resources so as to have FS (finish to start) only relationships. The results can never be correct when resource units were not consistent with the "planning unit". Therefore, first you need to decide which planning unit to use for shutdowns/turnarounds, and then tailor all your resources to be consistent with the established planning unit. As for sure you knew that some activities in shutdowns may only take an hour or even minutes.

Tom, P6 default leveling settings are awful. But there is an opportunity to select leveling criteria that may improve P6 schedule. I expect that LL that I never used uses leveling options that were set in P6. So I would suggest to try several criteria (late start, total float, duration, etc.) and select one that creates the best schedule for the current project. When data change other criteria may create better schedule and so the selection of best criteria shall be many times repeated.

As I understood Logic Leage analyzes but does not improve P6 schedule. Resource dependencies show why an activity is delayed. And I agree that it is useful (Spider Project also shows resource dependencies) for schedule analysis.

People will start to use automatic resource levelling when they will understand that it is reliable and creates better schedules than those created manually. Besides, it shall be required by management and this is the weakest point.

Tom Boyle
User is online Online
Joined: 28 Nov 2006
Posts: 258
Groups: None

Vladimir,

You've written nothing here for me to disagree with, though you've certainly explained it better.  My suggestion to investigate Logic League remains.  As I understand it, one click in LL... 

  • implements automatic leveling (using the P6 leveler),
  • infers resource links by examining the apparent resource logic decisions made by the leveler, and
  • substitutes the inferred resource links for the leveling delays and turns the leveler off.

The external database provides extensive coding of each logic relationship, and those inferred from the resource leveler are all clearly identified.  I believe different types of relationships can be enabled/disabled on the fly.

This would seem to support the benefits of heuristic resource leveling while also providing a clear logic basis for the scheduled dates.  I've never used LL (nor the other Schedule Analyzer software) so I can't vouch for its use in practice.  The developer, Ron Winter, is certainly one of the experts on the inner workings of P6.  The site linked by Raphael is the latest, though a bit dated (2003/P3).

The weak point would be the P6 leveler itself.  I have a sense that it won't get much use until it's improved, and it won't be improved until more people use it and demand improvements - Catch 22.

Rafael Davila
User offline. Last seen 1 hour 10 min ago. Offline
Joined: 1 Mar 2004
Posts: 4778

There are so many limitations within P6 resource leveling capabilities that in many cases the only option within P6 might be using manual leveling. 

  • P6 cannot level financial resources,
  • P6 cannot deal with automatic variable resource quantity leveling,
  • P6 cannot efficiently deal with multiple shifts on same activity,
  • P6 cannot deal with spatial resources,
  • and the list goes on. 

Because P6 is missing so much many might have no other option than to use archaic manual methods. 

No matter what method you use make sure you check the schedule is feasible and all constraints have been stisfied. 

Tom,

explicit crew flow logic that may be right at the project start will become obsolete very soon for many reasons. Creating new crew flow takes time especially in large projects. Shutdown projects require very fast management decisions and planners just do not have sufficient time for removing old and adding new links, playing what if, compare options, selecting the best sequences. Remember that they have a lot of other things to do because the model is constantly changing.

There are other problems with P6 and MSP capabilities: shutdown projects are usually done in several shifts, activity durations depend on assigned resource productivity, resources that have the same skills are interchangable, the number of workers that do the work may vary, etc. I expect that shutdown planners spend their time struggling with these and similar problems and adding manual leveling is too much.

So I think that resource leveling is the only option that is reasonable even though P6 resource constrained schedules may be improved.

Tom Boyle
User is online Online
Joined: 28 Nov 2006
Posts: 258
Groups: None

Mark,

  1. You are stuck with P6.  Thankfully, this is a very convenient reason to avoid engaging with Vladimir and Raphael on the relative merits of resource-constrained scheduling in P6 compared to others.
  2. User acceptance of automatic resource leveling in P6 seems quite low, with a general preference for using explicit crew flow logic instead.  Besides a number of other issues, the main reason for the low acceptance seems to be the incompatibility between resource leveling and Critical Path management techniques.  Total Float, Longest Path, and Multiple Float path calculations are all unreliable in resource-leveled schedules.
  3. While many construction projects are completed successfully using P6 with explicit crew flow logic instead of automatic leveling, I think you’ve correctly highlighted the limits of the approach in a turnaround project.
  4. I suggest you take up Raphael’s suggestion to investigate Logic League, the add-on for managing mixed technological, resource, and other precedence relationships in Primavera Schedules, using an external database.  The key claim is that Logic League is able to use P6’s resource leveler to automatically update explicit resource links in a complex schedule.  I know of others who have tried it unsuccessfully for similar purposes, though I don’t know the causes of their issues.  YMMV.

Good luck, tom

No Anoon, I do not understand what do you mean by linear calculations. Maybe some practical example of linear and non-linear calculations could help.

Maybe you mean something similar to scatter diagram that shows probabilities to meet several targets? In any case it is just one of many reports that can be created when the model was calculated. S-curve is one of many reports that can be also multi-dimensional. Please explain.

Anoon Iimos
User offline. Last seen 18 weeks 3 days ago. Offline
Joined: 22 Sep 2006
Posts: 1416
Vladimir, I'm not a mathematician (and I said I Guess), so pardon if I cannot explain it. Nonetheless, as far as I humbly know, number calculations (pairs) will eventually form a line as in a graph, like s-curve for example (please correct me if I'm wrong). Say you equate Scope with variable units and Time which may have variable units as well, plus perhaps you want it to arrive to a certain unit of currency, then perhaps you will form a Line. I believe you can expound further on this (please?)for the benefit of everybody. Thanks

Mark, you are right that manual leveling is not practical and besides it does not guarantee the best results.

It is pity that you cannot select your tools because Primavera has very limited resource scheduling capabilities required in shutdown projects.

Anoon, you wrote: "common software(s) are only capable of linear calculations." What do you mean by linear calculations?

Anoon Iimos
User offline. Last seen 18 weeks 3 days ago. Offline
Joined: 22 Sep 2006
Posts: 1416

Shutdowns / Turnaround works may need to be granulated as much as you can in order to be realistic. So its natural to have thousands of activities for perhaps simple works like tanks/vessels and appurtenances cleaning only (which may require hot tapping in some portions for continiuos operations).

I guess it is always how you structure your project activities down to realistic steps. And again, I guess your software is only capable of linear calculations when used automatically. Of course all activities which were not logically linked will fall in line exactly with the data date when pressing F9.

I'm always a fan of Mike Testro's Finish-to-Start only relationships (I believe this always minimize mistakes). When I guess we all know that common software(s) are only capable of linear calculations.

Mark Chapman
User offline. Last seen 43 weeks 1 day ago. Offline
Joined: 15 Oct 2018
Posts: 3
Groups: None

I missed a word in my initial post, the issue I 'm trying to discuss is manual resouce "leveling" vs automatic "shift F9" resource leveling.

1) Requires logically linking activities that have no physical link between them "Manual Resouce Leveling".

2) Automoatically spreading activities/resources to determin manning levels "Automatic Resouce Leveling". 

Using the second method once manning levels are calculated according to set limits, "F9" will return all available work to start on the data date.

I personally am against linking activities that have no physical constraint to work starting, that way there can be no confusion in a 50,000 activiy schedule as to when work is scheduled, why it can't be started and when it is avaiable for work.

Manual resource leveling can give the misrepresentation that some work scopes are being delayed due to other works, which can be a complete misunderstanding. If you have a 50 activities in a schedule and 12 months to complete this may not be an issue, however updating around 2,500 activities a day in a 16hr split shift with 6 planners over 20 days can be problematic.

However I'm not closed to others points of view, can manual resource leveling be successfully used and tracked during a large shutdown of around 200,000 hrs over 20 days?

We have Primaver P6 to use as our software, no other software will be considered this is out of my hands.

Rafael Davila
User offline. Last seen 1 hour 10 min ago. Offline
Joined: 1 Mar 2004
Posts: 4778

If your software is not capable of showing the resource dependencies then consider external software that will disclose such essential information.

Primavera P6 will not show you the data out of the box, if using P6 consider using Logic League.

Logic League™

Resource Links

Do not expect Logic League to improve on P6 resource modeling capabilities or leveling algorithm.

For software such as Spider Project it is easy, just click the mouse to see the links table where you should be able to see permanent logic links as well as temporal resource links [resource dependencies] duly identified as such. 

Non leveled schedule:

11111

Leveled schedule for an availability of five resources A showing resource liks using dashed lines:
22222

Essential Links Table showing and identifying Resource Links/Dependency:

33333
 

Anoon Iimos
User offline. Last seen 18 weeks 3 days ago. Offline
Joined: 22 Sep 2006
Posts: 1416

It is not Good or Bad but of course Necessary.

When you link and or sequenced activities, of course you will need corresponding resources to realize them.