(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.
when we start a project , one of the major requirement is to agree with the consultant the relation between activities , or methodology of how activities will be linked together in each field ( engineering, procurements , construction... ect), always all planners care about how to draw the relation , and not how this relation will feed me back during the stages of updating.<?xml:namespace prefix = o />
for example let us to study the relation between submission and approval for the engineering Items, of Corse first condition which control this relation is the agreed time of approval from the client which is mentioned in the general condition of contract , but how the relation will be :
Start to start with lag ( ex 30 day for time approval) or
Finish to start with negative lag or
Finish to finish with lag..
for the first wile it seems that the relation start to start with lag for the overlapped activities most accepted, because it’s give the owner the actual satiation in simple way, but this relation is not useful when the project is in delay and the contractor want to gain a progress in the engineering activities in order to recover the delay, because :
Whatever the contractor gain a progress in this activity ,this progress will not recover part of the delay. why?
because you use the relation start to start , which meen that activity of approval will start after activity of submission by 30 day , so if the contractor started to submit late, this will shift also the planned start of approval , and if the contractor start to recover this delay for this activity(submission) he will speed up the progress of submission on order to increase the percentage of progress in this activity( submission) , but unfortunately ,this will not affect the delay of the approval activity(means that planned start date and finish date of the approval activities will not shifted to be earlier) according to the progress gained .
from the above example, i want to highlight that when you use the relation start to start and finish to finish more in the program of work (base line) its will tying the activities and it will be difficult to you to recover the delay by speeding up the progress, because the progress (in the same actives) will not affect in delay when you have series of activities linked together by relation start to start, only you have one option which you can use to recover the delay, is to start the activity early more than planned or to finish the activity early more than planned, so the actual start dated and actual finish dates are controlling the delay only .
The best solution in order to have a good link between the over lapped activities is to link it by relation F.S with negative lag, this will give good results and will reflect on covering the delay when contractor have a progress in the submission. By explaining to the owner he will accept that for sure.
take care about the opinion of people regarding the negative lag , don’t believe when someone say that it’s not good,.
Printer-friendly version
Send to friend
THANKS HAITHAM.
Khalid,
I read the debate very well. I guess you did not read it well and just took the part that you want and left the conclusion which was clearly not favouring using that kind of logic.
If you believe that this way is perfect for you, go ahead and continue using it but do not promote for it because everyone here is trying to tell you that it is not a good practice but you're not listening.
Cheers,
Haitham.
Khalid,
I will answer in two different post, because I have an appointment soon, and I do not have enough time
Let me cite
I sad
"So imagine in your situation A (10days) ---------FS -5 --------> B (2 days) B will finish before A, that is if a is earthwork, b is pipe laying, in some places you will lay the pipe on the surface of the road, 'cause there is no ditch".
Your answer was
"I will answer you that also imagine that A (10days) ---------SS +5 --------> B (2 days) B<?xml:namespace prefix = o />
So B will finish before A also, so problem is not because you use the negative lag ."
My comment:
I never sad that SS relation is enough to describe overlapping. With this I wanted to show, that if you apply FS-5 it can result such an interesting situations.
I sad
1) cut into small pieces and apply FS
2) next but not so perfect solution is to use SS, and FF in this case B will finish after A even if it has started, because there is a logical dependency between the finish of B and the finish of A
But the most important thing to answer the followng question. Is this true that in case of such an overlappings B can be finished after the finish of A? Please answer this question for me. If your answer will be yes, than my next question: Then why don't you insert this dependency into the network logic?
I will continue with the next part, but first, please answer my first and second question
Got to go
Regards
Miklós
MIKLOS:
again i will replay your two point :
you said :
"So imagine in your situation A (10days) ---------FS -5 --------> B (2 days) B will finish before A, that is if a is earthwork, b is pipe laying, in some places you will lay the pipe on the surface of the road, 'cause there is no ditch".
I will answer you that also imagine that A (10days) ---------SS +5 --------> B (2 days) B<?xml:namespace prefix = o />
So B will finish before A also, so problem is not because you use the negative lag .
--------------------------------------------------------------------------------------------------------------------------
You said :
"Or you can imagine that after starting B, activity A will delay a year as I explained in one of my previous letter"
i will answer you that if A delayed 1 years, that’s mean that logic is changes, theoretically they didn’t become an overlapped activities, also answer yourself how practically you assume that A will delay and B will not delay and you put the concept from fist that ( B must be after A always, or B depend on A always)
Please run this in the program,
That after start of B , A is delayed 1 year
You will see that the remaining of B have the same start of the remaining of A
Thanks so much.
Khalid,
i was thinking how coulld i make my point more clearer for you.
i teach that only existing relations and not their mathematical substitution have to pu into the schedule.
what does it mean? let1s suppose that we have two activities A- foundation 2 weeks; B - walls (3 weeks)
the right relation A ----- FS say 1 week ------> B,
mathematically xou can convert it into A ------- SS 3 weeks ------>B
It gives the same result.
there is a simply method to decide wether a relation is corrector enough between two activities or not. you have to simply change the durations and check that what you see is what you want. for example change the duration of A into 6 weeks
Due to the SS relation B still will start on week 3 and finish on week 6. This is definitely impossible, so the relation is not good or not enough. a mathematical substitution will give the same result for the existing situation, but it won't work in case of any changes (eg, you change the duration of the following activity see below)
So imagine in your situation A (10days) ---------FS -5 --------> B (2 days) B will finish before A, that is if a is earthwork, b is pipe laying, in some places you will lay the pipe on the surface of the road, 'cause there is no ditch.
Or you can imagine that after starting B, activity A will delay a year as I explained in one of my previous letter.
All in all: in case of overlapping there is a dependency between the finish of A and the finish of B, or if you cut it into small pieaces ther is a dependancy between the last segment of B and A.
If this reasoning is acceptable for you, then you are convinced not to use negatív lag anymore.
Miklós
Hi Miklos:<?xml:namespace prefix = o />
Thanks for replay, but please do not judge (wrong) before you have the proof.
Considering you case # 4 , as you said , the both activities A&B are started and have actual dates , so activity B will not move according to the moved date of A (1/1/2012) .
if this what i understand from you , please tell me how you want an activity B to move (according to the movement of A and ) and it has actual date already. whict relation can apply this what you want<?xml:namespace prefix = u1 />
I answered all commenters, i give example ,
i wish really that i explained simply enough
Also i comment on your sentence
"This is a theoretical discussion so we cannot count on how Primavera or any other software behave"
by saying , primavera is one of the tools which you use to put on it your theoretical relation ,also its the tool which give the output to you, so you must to be aware how this tool is working with you inputs in order to maintain that this tool give you output similar to theoretical
Hi Haitham:
i appreciate what you said "Your reference is against your point of view", and i invite you to read this reference ( complete) carefully again.
Khalid,
Please do not get hurt, if I did hurt you it was without intention.
Miklós
To all those who comment without respect.
what the proplem of you to be out of limit.
you must to know, study and read more about how social networking with other nationals, before you write
this is for who comment without respect.
Mike,
My English was not enough, so I had to check my Bible :))
I've tried to find your thread you have suggested, bit i couldn't find it.
Where is it?
Miklós
Hi Miklos
Don't give up "kicking against the pricks". (Paul 26 : 14)
I have been advocating best planning practice on PP and other forums for many years.
If you would like some entertaining reading then have a look at my earlier PP thread on "Ban these planning abominations" and you will get some idea of the length and depth of ignorance that is endemic in the planning community.
They have no interest in improving their situation and I doubt if the new PP accreditation scheme will change things - just confirm and reward all the old bad habits.
Keep in touch
Mike Testro
Khalid,
I still say that you are wrong. And I will prove is with case #4.
But first a general comment: This is a theoreical discussion so we can not count on how Primavera or any other software behave. Couple of years ago I've made a tremendous amount of money on the problem that P3 files were converted into Primavera Enterprise, and they never had the same project duration, and never had the same critical path, and they never had the same start and finish for 90 percent of the activities. The discrepancy in project duration was sometimes 6 month or more.
So I would suggest to stay in theoritical level if you don't mind.
Continuing your example . Case #4
Imagine that you start wit A as it planned, 1/1/2011 and start with B as it planned at 6/1/2011. On next day your subcontractor on A says that there was an Eathquake in japan, so he will finish A on 1/1/2012, that is almost a year later!!! . So you put it into your schedule, but this would move your B to the end of 2011. But you have already started it, so it won"t be moved!!!! It means that the original relation was not good, just it looked like as it was good.
I think it can be accepted all of you - even if we know that there is a mathematical model behind, and models never able to simulate the real life situations - that we should chose relations which model the real life with the smallest discrepency. Here I proved that the relation you have appkied does not exist, cause we started with B even if A never will finish.
Let me continue this reasoning...
We also say that logical relations has to be formulated on that way that if something is changing we have to change only this thing. In this case you should give the new duration for A. If you do this and leave B with the same duration, B will finish sooner than A. This means that probably a logic is missing from your network, and this can cause problems, cause it will show, that B can be finished before the finish of A. In extreme situation you have to change the duration of hundreds of activities, just because of the changing duration of A. Which is unnaceptable. Network logic must be as complete as it could be.
So if you admit that there is a logical dependency between the finish of A and the finish of B, than why don't you insert into your logic. The most precise way as Mike sad is to cut it into pieces and apply FS relations, the next is to use SS, and FF +lags between these activities. The worst is to use negatíve lags. It looks as it would good, but as it doesn't describe the right logic, this will cause problems later, during updating.
So this was an example for my statement # 1 about negatíve legs (you can read below) This logic do not exist.
I hope I was clear enough, and you will understand my point. If not, then I give up, l give back my professorship, and will retire. :)) (Sorry no offense intended)
All the best
Miklós
Considering the all cases which can be achieved from the negative Lag between A and B:
For easy way to draw the activities bar, i will consider this template to define my bar:
<---start date--activity name--activity at complete duration--end date--(successor relation, lag)--->
1) Original example (when data date =1/1/2011):
<---1/1/2011--A--10--10/1/2011--(F.S,-5)--->
<---6/1/2011--B--10--15/1/2011--(F.S,-5)--->
2) Case 1(when activity A finished on 8/1/2011, Activity B didn’t start):
<---1/1/2011--A--8--8/1/2011--(F.S,-5)--->
<---4/1/2011--B--10--13/1/2011--( , )--->
this up bars are wrong , activity B will not start on 4/1/2011 sure because the data date is 8/1/2011 , so the correct is that activity B will start on 8/1/2011 and the correct bar is :
<---1/1/2011--A--8--8/1/2011--(F.S,-5)--->
<---8/1/2011--B--10--17/1/2011--( , )--->
From case 1, you see that till now the logic is right and the relation look like that its changed to be F.S
also this case will enlarge the overall duration due to the delay of B ,but for your hint … this enlargement is due to change of arrangement of work,
if we consider A is( column rebar), B is( column formwork), so if A finished early as in example that’s mean the contractor finish all column rebar work and after that started formwork , so the logic is applied from this relation.
3) Case 2(when activity A didn’t start , Activity B started on 1/1/2011):
<---1/1/2011--A--10--10/1/2011--(F.S,-5)--->
<---1/1/2011--B--10--10/1/2011--( , )--->
From case 2, you see that till now the logic is right and the relation look like that its changed to be S.S, even if B started also before that date the relation also will be start to start because A will not start before data date.
this case will reduce the overall duration due to the delay of A ,but for your hint … this reduction is due to change of arrangement of work also,
if we consider A is( column rebar), B is( column formwork), so if B started early as in example that’s mean the contractor started column formwork parallel with rebar work , so the logic is applied from this relation also.
dont till me what a bout if you take the example of slabs ( form work and rebar) , i will till you put B as (rebar) and A as (form work) and study the example again.
at end I want to highlight that the you must to consider the data date and put it in to your account to study this relation, also you must to consider the behavior of primavera.
Khalid,
Your reference is contrary to your point of view. It states that "Networks utilizing finish-to-start relationships with a negative lag can absolutely provide a reasonable calculation for the completion date of a project. However, schedules using this approach to overlap work activities, by definition, eliminate many schedule activities that would otherwise be included within the CPM network. this elimination process also can eliminate the transparency of the project plan as well, and the resulting lack of transparency can substantially add to the project labor inefficiencies. This added inefficiencies can be much more expensive that the cost of updating a few more schedule activities per week or month.........."
That means that using that kind of logic is merely theoretical and does not work properly in the real world.
Khalid,
I went through the paper you've recommended. I'm still not convinced, all the examples were given there was an example for
1) this logic do not exist betwen these two tasks
2) logic exist but in an opposite way.
Just let me explain this through their first example given below
"Negative lags have been used to forecast when an activity will start prior to the finish of a precedent activity. An example of this is the start of check-out activities. A start up supervisor may say "I will start my check out two weeks prior to the project being complete" meaning that two weeks prior to when the project will obtain mechanical completion, he will mobilize his start up personnel to check out the facility."
Imagine that there is a Finish to Start - 2 weks, or FS-2 weeks or FF-2weeks (all the same if the finish of project is a milestone) The guy will start the checkings but within a week due to some reason (the contractor goes to bankrupcy) the project will be finished a year later. This means that checking should have been start a year later. But it has started. This means that this logic did not exist between the two activities. Funny thing if I understood their next sentence well, that in the next sentence they explain what would be a predecessor of check ups. They can start with check ups after finishing of mechanical works. And project can handed over, if check ups had been finished. If this is not the case (I did not understend well) than a real predecessor has to be find.
Anyway In this case the logic exist but in the opposite way. project finish depends on check ups, it can be completed after a two weks check up. This is a FS0 from check ups (wich is planned for two weeks) to the project delivery milestone.
So this was an example for reverse logic, which is very dangerous especially when tracking projects. Itt can resulted in situations when roof comes first and then the foundation.
Hope I was clear enough, and this can help to explain my points. It is easier for me near a blackboard.
Back to the original question to describe overlapping a more or less good solution is SS and FF relation at the same time but the most precise as Mike told us, to cut the whole stuff into small pieces and apply FS0 between the pieces within and between activities. But who has time and effort for this? So it is up to you, if the risk is small for delaying the activitis, or this is a draft then apply SS and FF to ensure overlapping.
All the best Miklos
Anoon,
Your example is a good one for the first point of my previous letter cited here
1. this logic is not exist between these two activities
The tru logic is FS7days from A to B (where 7days are consequtive that is calendar days and not working days).
(You can break if you make first)
I'm still waiting for justifying my option #3, which states that negative lag might describe an existing logic, but no one has found a proof for this so far :)) (at least not me)
Miklós
for all who afraid from using negative lag, read this :
http://www.calveyconsulting.com/files/PS02_Great_Negative_Lag_Debate.pdf
thanks for all .
Start-to-Start with Lag?
How about casting, curing and breaking samples?
1. Task A - Cast concrete sample
2. Task B - Break Sample (after 7 days)
A -SS7 - B
If A is delayed, will you allow to break sample after 5 days?
Hi Haitham,
As a lecturer we teach to our students when they check a plan, they have to check (amongst others) all the relations with negative time.
We say that there are three options:
So I agree with you. I would be grateful, if any of you could present a real example for the third case.
Dear Khalid,
Thank you for snapshot this is a good explanation. I am not fan of "Progress Override" but sometimes it is becoming usefull if you use dummy relations as FF+lag. I am suggesting ss+lag and ff+lag together with progress override due to facts you referenced. I will try to make my demonstrations like yours (Unfortunately I do not have enough free time for making posts like Mr.Davila). In this exapmle we can see the flaw of duration based CPM and float calculations. There are many factors as planning level and project type (eng., proc, const, EPC, pipeline, building etc.). There are 2 options in P3 (or in P5/6) for modelling projects that contains majorly overlapped activities works according to me:
1- Make your baseline as I mentioned (normal relations + ff+lag (dummy relation) in order to avoid wrong finish dates in progress override).
2- Use Retained logic and check all of your relations (especially out-of sequence and in near critical path) for every update. I think this option is the most healthy one. According to some contractual items or best practices of different companies; you cannot change relations in each update or the changes are being subject to client approval. This will result in modifying relations and out-of sequence activities only once in log periods or making weekly updates nearly impossible.
The best solution to this situation is Resource Based Scheduling where resource levelling of majority softwares cannot result 100 % correct. You should check Spider(c) software. You can in the Spider item in forum. I hope they will have bigger portion in market.
Regards,
Safak
Hi Mike,
Can you please elaborate, how is it possible to never use FF, SS or SF links with or without lead lags?
In the construction industry. How can you put FS links to all the activities while there are a lot of overlapping activities??!!
using a simple example, rebar and shutter for columns are always overlapping or should i wait for all the rebar to finish then fix the shutter or maybe i should make an activity for each column to use only FS links??
In my opinion the best way i found was using both SS+lag & FF+lag
and regarding using -ve lag, no i didn't find that useful and its giving false indication. For me its like using constraints.
I have not seen any practice recommending using -ve lag and at the same time i don't recommend using progress override unless its necessarily required.
Hi Safak
i always proof my all Openions , please reed this book which define how if you use "progress override" for controling the ralations of progressed activities .
Hi Mike,
Engineering approval logic is same as a production line. You should keep production rate of the steps close in order to avoid stack ups of documents in the one end. I agree with you; for construction trades at site (except high-rise b. core construction or a long pipeline) ladder model is not suitable.
Regards,
Safak
Hi Safak
You describe the Ladder method which is well suited to prodution line process where Just In Time logic is natural.
It is not well suited to the construction environment where interlocking trades have different production rates.
You will end up with the wrong resource balances and unecomical deployment.
Best regards
Mike Testro
PP Members,
Overlapping activities as the example are always subject to bugs of CPM method. LoB or other studies can provide better modelling in this options.
The best model I have observed in CPM is (SS+lag and FF+lag) with "progress override" in this kind of situations according to me.
Regards,
Safak
Hi Khalid
I agree with what you say and put it a bit more succinctly my code of best planning practice includes:
Only use FS links between tasks - NEVER US FF SS or SF links with or without lead lags.
Why?
Because you will never get a true dynamic critical path otherwise.
Best regards
Mike Testro
Brother,
Could you explain me more clearly ?
Regards,
Saud
Email: PPadmin@planningplanet.com
Replies