New Spider Project Functionalities
In the Open Standard
Rafael DavilaMember for 22 years 7 months
- Created
- 88
- 7
0
Vladimir:
At times new functionalities are added or improved by Spider Development Team. In order to know and understand them I suggest using this topic as to keep us informed and to discuss how these are to be used.
Every time you post a new functionality we will get an e-mail from PP and many Spider Project users in the English speaking world will get first hand notice from Planning Planet. I hope others eventually will find how great this place is and will join us.
I just downloaded the last update, version 10.01.24 date 10.02.2010. The very first thing I did was I looked for the menus and also for the available fields. Today I noticed a field I have not noticed/seen before and seems like something I was looking long ago but prefer you to explain me what it means. I am referring to the field named Code in Original Project, is it a baseline, the very first saved code? How can I re-set/fix it so it represents the officially approved original schedule? What if I add activities, how new activities codes are tracked? Does it represent a unique ID code no one can edit? Is it created when from version 1 you create a new version ?
Best regards,
Rafael
At times new functionalities are added or improved by Spider Development Team. In order to know and understand them I suggest using this topic as to keep us informed and to discuss how these are to be used.
Every time you post a new functionality we will get an e-mail from PP and many Spider Project users in the English speaking world will get first hand notice from Planning Planet. I hope others eventually will find how great this place is and will join us.
I just downloaded the last update, version 10.01.24 date 10.02.2010. The very first thing I did was I looked for the menus and also for the available fields. Today I noticed a field I have not noticed/seen before and seems like something I was looking long ago but prefer you to explain me what it means. I am referring to the field named Code in Original Project, is it a baseline, the very first saved code? How can I re-set/fix it so it represents the officially approved original schedule? What if I add activities, how new activities codes are tracked? Does it represent a unique ID code no one can edit? Is it created when from version 1 you create a new version ?
Best regards,
Rafael
88 replies
Oldest first- 0Vladimir Liberzon ·Rafael,
thank you for the proposal. I will do it.
Code in Original Project is not what you expected.
When projects are consolidated in the project portfolio activity codes in different projects may be the same. Spider Project can open the portfolio as one large project and do with it everything that can be done with the project model (scheduling, levelling, any kind of calculations and reports and some additional). So there is a need for unique activity codes in the project portfolio.
These codes are assigned by the software but Spider remembers original activity codes. After portfolio calculations and analysis separate project models are sent back to their managers. They shall receive new versions of their project models with the original activity codes to be able to compare them with the previous versions, etc. So Spider Project keeps two kind of activity codes - project and portfolio.
We recommend to keep activity codes the same in all project versions. This is the key field for comparing different project versions, EV, trend analysis. If project scope was changed and new set of activities appeared we suggest to make changes not only in the current project version but also to create new baseline. The history of baselines is kept the same way as Spider keeps project archives (all previous project versions are stored as the separate files).
Best Regards,
Vladimir - 0Vladimir Liberzon ·New release of Spider Project includes two new features:
1) Calculation of activity resource constrained super floats that show when activity can finish without delaying project finish if their execution will be done slower than expected (with splits, etc.).
2) Now Spider Project users can place any number of diagrams in the window under Gantt Charts. This windows is resizable, has scrolling, and you can move any selected diagram up and down the window.
Hi Rafael,
you asked for these functions. Now you can try them.
Will keep you informed on future developments.
Best Regards,
Vladimir - 0Rafael Davila ·Vladimir,
You are late with regard to the unlimited number of diagrams, I have been enjoying it science yesterday, a few more (unlimited) than the single diagram available in SureTrak under Gantt Charts.
About superfloats just got them in today’s update. Looks beautiful, goes directly to what I was looking.
Even when I already got the most important functionalities such an unsurpassed resource leveling and resource assignment functionalities plus true resource constrained float these refinements are welcomed. At times they become very useful, some help you to communicate the ideas, some help you understand and tame PDM logic.
Thank you, very much.
Best regards,
Rafael
- 0Vladimir Liberzon ·In the new version of Spider Project it is possible to make resource dependencies visible.
The software creates artificial links that are shown by dotted arrows. If an activity is delayed because the assigned resource is not available and starts when the resource finished to perform another activity these activity become connected by resource dependency.
It makes resource constrained schedule analysis much easier.
You can select to show or to hide these links, or to show only these links hiding regular links.
Resource Critical Path is now visible as a path and not just a sequence of resource critical activities.
Another new functionality: you can use Status Date, Data Date, Target Date, etc. in formulae.
Best Regards,
Vladimir - 0Rafael Davila ·Vladimir,
I have used the functionality for creating links quite often as soon as it became available. I noted that if you filter for predecessors of an activity when all predecessors are resource dependencies if you select create links the software can see them but if you do not select for the software to create the links then it cannot see them.
I wonder if you know if P6 longest path filter can identify longest path when it is determined by resource dependencies. Can you tell me? I do not have P6 and wonder about the filter I believe was not available under prior version P3e I used before.
I am including a sample job I would like you to tell me how the longest path to the last activity would be displayed by P6 filter, I suspect it will not be able to identify full path as to include resource dependencies. Prior to resource leveling all activities are scheduled to happen at the same time.
Sample Job with no activities linked by logic, all links are resource dependencies shown by dotted lines:
The following view is better than using graphic display of links, I frequently get lost when links span several pages. It shows all links are resource dependence links. If you unlevel the schedule then all links dissapear and the table will be empty in this sample job.
Best Regards,
Rafael - 0Rodel Marasigan ·Hi Rafael,
P6 don’t have functions of resource link to identify predecessor and successors of resource dependency therefore P6 don’t have a capability of filtering full path of non logic activities. It will show only the last activity but enable to filter critical path created by resource.
By the way, how to create resource link using resource dependency in spider project?
Best Regards,
Rodel - 0Rafael Davila ·Rodel,
Go to Calculations/Resource Constrained Scheduling Options and mark Create Resource Dependencies. Perhaps is not a default option as for large jobs it adds to the computing time. Also go to Gantt/Options and make sure Hide Resource Dependencies is unmarked.
There is also a filter for All Preceding Activities you should learn to use, is available when you select an activity line and right click. By the way in Spider to move activities you use cut and paste no drag and drop, this prevents errors but is not what we are used to do with Primavera products.
Because of the time difference many of your questions I will reply next day but always will.
Best regards,
Rafael - 0Rafael Davila ·Rodel,
For the moment Spider Project do not have a longest path filter per-se, the filter is a filter for all preceding (and linked) activities, it will show preceding activities only if links are visible, so if there are links that are dependencies links it will only display them if you have selected for the software to create dependencies links.
I had a previous conversation with Vladimir about longest path but he did not wanted to create a limited filter that would not filter longest path in the presence of resource dependencies and even further he want it to consider other constraints such as financial constraints. I believe the creation of resource dependencies links can be a step to it but because these can appear and disappear as the schedule moves there is the question of how useful it is.
Best Regards,
Rafael - 0Rafael Davila ·Rodel,
Filter is to preceding activities; it will show all paths leading to the selected activity and filter out all other paths, therefore after having created all links it might be a matter of doing the PDM calculations to the isolated paths to get local critical path and floats.
Saving the Schedule under a different name and then identify preceding paths to delete unrelated paths is not very practical because as soon as you make any changes to the schedule new possible resource links might not be available as these where filtered out. For it to be practical under resource constraining I believe it must be done by the software without eliminating other paths.
Spider for long time calculated Resource Critical Path without disclosing the created links, now as added functionality it can display the resource dependencies links.
Best Regards,
Rafael - 0Rafael Davila ·Vladimir,
As per your posting #6:
"New version of Spider Project supports not only formulae that link and calculate project fields (columns) but also formulae that link and calculate cells (or the sets of cells) of different project tables."
Can you expand on how to get access to these fields, perhaps some coding is needed. I particularly noted the issue on reference to sets of cells this is kind of something I was hoping for long time, I would like to make conditional statements based on values on other fields, at times within the same table line, at times within a set of cells within the same table, but looking at different tables is kind of a welcomed surprise.
Best regards,
Rafael - 0Vladimir Liberzon ·Rafael,
an example: DurPlanD [Code, 1020] = 5
If this formula will be executed the duration of activity with Code 1020 will become 5 days.
Another example: Calen [Сode, 1020] = Calen [Code, А, Resource] means that Calendar of activity with code 1020 will become equal to the calendar of resource with code A that is taken from the table Resource.
Special functions:
#CurrentDate - Current Date
#DataDate - Project Data Date
#TargetFinish - Project Target Finish
#ProjCode - Project Code
#ProjVersion - Project Version
#ProjName - Project Name
Full description in English will be published soon.
Best Regards,
Vladimir - 0Rafael Davila ·Vladimir,
Thanks I got it, I even tried to make an illegal move by changing the duration of a productivity type activity. The software did not apply the change and protested, similar to when trying to apply non valid changes to database tables when importing values from an external table. I will be impatiently waiting for the full English description as not to miss any detail.
Best Regards,
Rafael - 0Stephen Devaux ·Hi, Vladimir!
How are you? A long time since I saw you in Boston! You may have noticed, I haven’t even been on the PP site for many, many months! On top of everything else (including both teaching and studying at the university level!), I have had a number of significant health issues. (Anyone who has ever had a kidney stone will appreciate what it might be like to have one for two months!) But I’m feeling much better now, though 13 kgs. lighter than when we last met.
Some questions:
1. Has the DRAG functionality been implemented fully in Spider? If yes, has there been any reaction to it?
2. Does Spider have a user base in Germany? (I’m working with a German client who might, down the road, be interested.)
Hope everything is good for you.
Fraternally in project management,
Steve the Bajan - 0Vladimir Liberzon ·Hi Steve,
I was worried about you and asked about you the people whom we know both. I am glad that you are fine and resume your activities.
DRAG is in Spider Project but we did not proceed with the cost issues. People asked what it is and how it may be used. I suggest to start the discussion on DRAG here - I expect that some of our users will participate in the discussion, others may be interested in the concept.
We have few customers in Germany. In particular you may mention Thyssen Shachtbau.
Best Regards,
Vladimir - 0Rafael Davila ·Vladimir
BEFORE RESOURCE LEVELING ACTIVITY 6 DRAG =1
AFTER RESOURCE LEVELING ACTIVITY 6 DRAG = 1? PERHAPS AFTER CREATION OF LINK DEPENDENCIES DRAG COULD BE RECOMPUTED? SHOULDN’T IT BE 2 DAYS FOR ACTIVITY 6 AFTER RESOURCE LEVELING?
FOR THE FOLLOWING SAMPLE I USED STANDARD RESOURCE LEVELING AS OPTIMUM WOULD GIVE MINIMUM DURATION OF 7 DAYS.
Best Regards,
Rafael
P.s. Did you make some improvements on resource leveling algorithms on last update? - 0Vladimir Liberzon ·Rafael,
we do not recalculate DRAG when level resources.
We may try after introducing resource links.
Resource levelling was not changed yet though we work on some improvements on material levelling.
Thank you for noticing a problem with restoring risk distribution curve, we will make improvements soon.
Best Regards,
Vladimir - 0Rafael Davila ·Vladimir,
I know determining best curve fitting depending on spread of input data for you will be a challenge you can solve with excellent results, but remember about the Expected Value or Mean, in case of modeling Liquidated Damages this single value will give you some sense on what direction and how much the distribution function is moving, easier and more intuitive than comparing distribution curves.
About DRAG after resource leveling this can be more complicated as with resource leveling source dependencies can shift, if it can be computed, and links are needed then perhaps Spider Project will be unique about DRAG after resource leveling. I don’t know how or if if it can be computed using resource leveling but the easy way is to challenge you.
Best regards,
Rafael - 0Stephen Devaux ·Hi, guys. A few comments about DRAG:
First, a while back now, for the first time while teaching a class, I demonstrated how to go in and optimize/recover a schedule using software to compute DRAG. (Since the company uses MS Project, I used the Sumatra add-on to MS Project.) It went over GREAT! I think actually showing people how to use is very helpful and persuasive. The customers I’ve showed it to since (and I taught a class to a different engineering company a month ago when I was over my health issue and where it once again went over great) have just been blown away by it. As I get back into teaching again, I intend to incorporate it regularly.
At my suggestion, Sumatra is developing for their website a short "case study" to demo how to use it to pull in a schedule. I would recommend that Spider would do the same thing (it’s really an education issue), and I’d be happy to help with developing that, if you are interested.
Second, resource schedule DRAG is a result of two different things: CPM schedule DRAG as well as resource availability DRAG (or from my book Total Project Control, RAD) which can change the critical path. It’s very valuable to separate out the delays caused by resource availability (and its cost), as that is the way we can go about justifying the needed resources to make the resource availability DRAG go away.
Fraternally in project management,
Steve the Bajan
- 0Rafael Davila ·Stephen,
We as a contractor after we develop the initial schedule the main issue become on how to recover our schedule, almost always something happens that delays the schedule and we have to recover. Perhaps the saying that activities will expand duration to available time is true; the Contractor usually waste saved time and then got to recover once again.
For months I kept asking what is driving my schedule, what is driving my schedule after resource leveling, once and again and again. Maybe Vladimir developed the functionality to display resource dependencies to get rid of me. Well now I want to know who is next into driving my schedule, or how to avoid the driver. Perhaps Longest Path theory (and second to longest) can be of help here, but if resource constraining is not taken into account it would fall kind of short to me.
Much is being said about good scheduling practice but few about the tricks on how to tame the schedule.
Best regards,
Rafael - 0Stephen Devaux ·"Much is being said about good scheduling practice but few about the tricks on how to tame the schedule"
Rafael, I love your expression: "(T)he tricks on how to tame the schedule"!
Here would be the functionality I would see as perhaps solving your problem. (Maybe it already exists in Spider, or maybe Vladimir can incorporate it easily.) Would a report that does the following help resolve your issues?
1. List all activities on the (new) longest path, from earliest dates to latest.
2. Subtract [start date of each activity on original schedule longest path] from [start date on each activity on new schedule longest path]. Any such activity with a value > 0 has been delayed by an earlier activity/resource/constraint.
3. Display/highlight activities where the value computed in #2 is different in Successor Y than in Predecessor X, as it indicates that the delay is due to something specific in Y (after X finishes).
4. Display DRAG and assigned resources of Y-type activities, look for the cause (which will probably be due to resource constraints), and use DRAG computation to try to recover.
5. Subtract [difference between start and finish date of each activity on original schedule] from [difference between start and finish date of each activity on new schedule]. Any such activity with a value > 0 has been delayed by some aspect within itself (resource availability DRAG?), and now has a longer duration than it did.
6. Display DRAG and assigned resources of all such activities (which now had a longer duration), look for the cause (which will probably be due to resource constraints), and use DRAG computation to try to recover.
I think this should allow very quick identification of the causes of delay, and the DRAG computation would help direct the user to where to go to reduce the impact.
It seems to me that the above process would also be of huge value in forensic analysis, allowing the user to identify the precise locations of activity delay on the as built critical path (ABCP) and their DRAGs versus DRAGs on the baseline schedule critical path.
I’m curious both as to whether you think this would help, Rafael, and as to whether Spider (which really seems to have more useful scheduling functionality than any other product out there) either can do this now, or would be able to with just a tiny bit of tweaking.
Fraternally in project management,
Steve the Bajan
- 0Rafael Davila ·Stephen,
I do not know of longest path functionality in Spider Project, therefore I could not follow your suggestion within Spider. In the following Spider Project example longest path is Activity 1 – Activity 2 – Activity 3 while critical path is discontinuous Activity 2 – Activity 3 (Activity 1 excluded as it has float).
Although I do not have P6, through PP debates I could figure out P6 misses longest path driven by resource leveling. Therefore it is kind of useless on my resource driven schedules.
Compressing the schedule have some cost, that it is a time-cost trade off issue, even if using some weightage to measure resistance to compression for each individual activity. I always believed that the use of time-cost trade off routines would be the key to determine best mathematical compression strategy. I had the idea that efficient use of DRAG was limited to the case when required network compression is not substantial but that it can be combined with some time-cost consideration. Yes near optimal solutions are acceptable, but remember they become far from near at some point.
For the purpose of this discussion I am interested in Drag as an alternative strategy to get solutions with some consideration on cost and not merely duration.
Just consider the simple case of the above 3 activities sample job projected to take 15 work days. Say we need to find the best way to reduce the schedule to 12 days under the following conditions. We would also be interested on knowing the cost even if optimum as perhaps liquidated damages might be less expensive at some point.
Reduce duration of activity 1 – one day 300$, 2nd day 250$ and 3rd day 250$ (max reduction)
Reduce duration of activity 2 – one day 325$ and 2nd day 350$ (max reduction)
Reduce duration of activity 3 – one day 200$, 2nd day 250$ and 3rd day 350$ (max reduction)
The solution would be obvious because they are all in tandem but in case of hundreds of activities the procedure could create multiple parallel paths as DRAG is being consumed.
Of course true time-cost trade off functionality will also tell you about when it make sense to increase duration of some activities with enough float when increasing duration reduces the activity cost.
Best regards,
Rafael - 0Stephen Devaux ·Hi, Raphael.
Your post raises a whole bunch of issues. It would take most of a book to address them all - but I already wrote it: Total Project Control from John Wiley & Sons. If you want an abbreviated description of the TPC approach, try ProjectsatWork.com and do a search on my name for a series of articles.
Let me try to help with some of your points, though:
"Compressing the schedule have some cost, that it is a time-cost trade off issue, even if using some weightage to measure resistance to compression for each individual activity... I had the idea that efficient use of DRAG was limited to the case when required network compression is not substantial but that it can be combined with some time-cost consideration."
DRAG is simply the amount of time an activity is adding to the project duration. I’d suggest that this is hugely important info for any planner to know about their project. How they use it is another matter: time/cost trade-off; time/scope trade-off; project recovery from slippage; resource justification; resource targeting; improving organizational staffing levels; even postmortem ABCP and forensic analysis.
Again, DRAG is just important critical path data; how you use it is optional.
"For the purpose of this discussion I am interested in Drag as an alternative strategy to get solutions with some consideration on cost and not merely duration."
Absolutely! The key is that time is money, often huge amounts. How much is it worth? That depends on the project and on the contract. On a cost plus contract with no time incentives or penalties, compressing the schedule can actually be disadvantageous to the contractor. (That’s one reason why no customer should ever accept a cost-plus contract without time clauses.) And a contractor who has a great scheduler should seek such clauses as an opportunity to add profit.
What is the value of pulling in a schedule? On projects I consult on (nuke outages, DoD/aerospace, pharma), the cost of time is huge! Adding $10,000 to reduce an activity’s DRAG by five days can result in $90,000 or more in profit. (The number that was being tossed around in pharma eight years ago was $32... per second! In general, if the value of pulling in the schedule is estimated at only $10,000/week, and it’s going to cost $8,000 to reduce the DRAG by a week, that’s pretty close to a wash. But there’s lots of occasions when the value of reducing an activity’s DRAG is dramatic. Especially during recovery.
Fraternally in project management,
Steve the Bajan
- 0Rafael Davila ·Stephen,
No doubt DRAG keeps you focused on the Critical Path. I guess from now on I will keep it on my mind, just want to make the best use of this simple metrics many have not paid due attention, well at least me, until now that you brought it to the table.
About saving 90,000$/five days it might be even more in case of a mega-drug with annual sales over a billion dollars, this give me 2.73 million per day in sales x 5 = 13.65 million in sales lost forever to the generic manufacturers once patent expires. The profit got to be a fair percentage of sales. Not everybody is on the opulence; most of us must find the most economical path, brute force at any cost is not an option.
When you get into more than a single critical path, crashing become more complicated, it becomes a linear programming issue, see the following reference for a birdseye view on the issue.
http://people.brunel.ac.uk/~mastjjb/jeb/or/netcpm.html
One of the assumptions of the model is that it assumes there are always enough resources available for crashing.
Best regards,
Rafael - 0Rafael Davila ·Stephen,
From:
http://www.stat.mq.edu.au/Stats_docs/stat321/Wk6Handout1.pdf
http://www.stat.mq.edu.au/Stats_docs/stat321/Wk6Handout2.pdf
Criteria for Selecting Activities to Crash:
There are several characteristics that mark or highlight an activity that exists on the Critical Path as a better candidate for crashing.
1. Must be on the Critical Path. Crashing noncritical activities that already have slack only buys more slack and doesn’t shorten the project duration. Only critical path activities drive the project and crashing them will shorten the project duration.
2. Precedes multiple activities. When an activity bottlenecks numerous succeeding activities, it is a great candidate to shorten. Once this activity is shortened, it allows the multiple activities to begin.
3. Long duration. An activity that has a long duration offers more potential time gain from crashing it.
4. Lower cost per period gained. Activities that cost less to crash are preferred. These include those requiring lower paid, lower skilled workers or other resources that are otherwise sitting idle.
5. Early in the project (the Sunshine Rule). If you fail in crashing the activity and it takes longer than planned, it is still early in the project. Thus you still have recovery time. Also, typically demand on resources early in the project is lower than other times, and they should be readily available.
6. Labor-intensive. When an activity is low skill labor intensive, it is easy to add people to help complete the project early. When an activity requires high skills to complete, it may be hard to find qualified individuals who are capable of completing the task.
7. Subject to common problems. Try to pick activities that are subject to higher probability of common problems. Shortening the duration lowers the exposure time and lessens the chances of having a problem.
Steps in Project Crashing:
1. Compute the crash cost per time period. If crash costs are linear over time:
Crash cost per period = (Crash cost – Normal cost) / (Normal time – Crash time)
2. Using current activity times find the critical path and identify the critical activities
3. If there is only one critical path, then select the activity on this critical path that (a) can still be crashed, and (b) has the smallest crash cost per period. If there is more than one critical path, then select one activity from each critical path such that
(a) Each selected activity can still be crashed, and
(b) The total crash cost of all selected activities is the smallest. Note that a single activity may be common to more than one critical path.
4. Update all activity times.
5. Cease crashing when
– The target completion time is reached
– The crash cost exceeds the penalty cost
If not, return to Step 2.
To the above I would add:
If optimum crashing results in crashing late activities you can factor late activities cost to crash as to artificially increase the crash cost and use this factored value on the above procedure as to obtain alternate crashing strategy. Then compare unfactored estimated costs to determine if higher cost of early crashing is justified.
DRAG metrics will tell you on the spot how much you can reduce the duration of an activity before a new additional critical path must be considered.
Stephen, please I would like to hear your comments about this procedure as until we do not have a crashing routine embedded into our software we must somehow do our best.
Best Regards,
Rafael - 0Stephen Devaux ·Hi, Raphael.
Normally, I’d be reluctant to criticize someone’s theoretical advice as long as there is some that’s wise (and there certainly is in what you’ve quoted). But since you’ve asked for my comments, here goes:
The main thing is that the authors aren’t aware of DRAG. It’s not their fault, because the concept only became public knowledge when my book was published (1999). And since only three PM packages (including Spider) compute it, unless they’ve read my book or one of my articles, or sat through one of my courses, they could hardly be aware of it.
But DRAG really changes the whole schedule-compression ballgame -- it’s by far the most important concept for compressing a schedule. All the other considerations mentioned in your post may be relevant -- but number one is DRAG. For instance, let’s take this one:
"3. Long duration. An activity that has a long duration offers more potential time gain from crashing it."
Long duration is meaningless without the DRAG concept. You can have a 25D task on the CP and its DRAG might only be one day. Conversely, you could have a 5D task on the CP
and its DRAG might be 5D. The latter offers the opportunity to gain 500% more time despite having only 20% of the duration.
A little over a year ago, I taught a class at a multinational DoD/aerospace contractor. When I explained DRAG, one engineer reacted dramatically. She explained that a year earlier, her team had been tasked by the client with pulling in by six weeks a mid-project task that had, from the total 5,000-activity programme perspective, over 200D of total float. The whole team flew to San Diego and met on a Monday morning. Most already had suggestions for activities they thought could be compressed. After about seven hours of tweaking, they gave the changes to their scheduler to input to the software -- and the schedule came in by one day!
After adjourning to the nearest bar, they came back and worked the rest of the week, just randomly trying things. As she put it, the concept of DRAG would probably have saved them eight person-weeks in San Diego, not to mention the airfare.
DRAG is NOT the total answer -- but it provides an "Occam’s razor" for selecting the best candidates for compression. Obviously, there are other considerations. One is resource elasticity: how much can an activity be compressed if resources are added? The TPC metric is DRED, or doubled resource estimated duration: how long would an activity take if resources were doubled? With doubled resources, a 20D activity may become 10D ("perfectly resource elastic"), or 13D (more likely!), or 17D, or remain 20D, or expand to 30D (e.g., building an avionics dashboard in the cockpit of a jet fighter).
Having this second estimate, based on doubling resources, provides the basis for the conversation when it turns out that a 20D task with 15D of DRAG has a DRED of 12D. We know from the DRED that the work is quite resource elastic. Now we need to find out:
(1) Do we have to double ALL its resources or just specific skills?
(2) If we added even more resources, would the duration keep shrinking until its DRAG (and DRAG Cost) disappeared?
(3) How much would the resource(s) cost?
Raphael, the subject (as you clearly understand!) is a complicated one. Since you seem so interested, I really would urge you to pick up a copy of my book -- I think you’d find lots of nuggets that would interest you.
One more item that’s crucial: you’ve got to know the cost of time on the project. Without that info, NO increase in resources (and no crashing or fast-tracking) can ever be justified because resources cost money and, if time doesn’t, why would you ever pay to save it? (Of course the truth is, time on the CP (DRAG) almost always costs more than the resources that would save it!)
Finally, I know you have Spider (which, from what I know of it is a great package!). But if you also have a copy of MS Project 2007 available, I urge you to download and check out the trial version of the Project Optimizer available from Sumatra.com. Not because MSP is anywhere close to as good a product as Spider (it isn’t!), but because the Sumatra folks have included a walkthrough document of how to use DRAG to optimize a schedule on a sample project file. I think that might answer a lot of your questions and you could then use the same techniques in Spider.
Hope this helps.
Fraternally in project management,
Steve the Bajan - 0Rafael Davila ·Stephen,
In the absence of sophisticated time cost trade off embedded in my software of choice I believe DRAG is quite important, otherwise you will miss the exact point when multiple paths emerge.
I already asked for the demo software you proposed, who knows maybe it will give Vladimir some ideas and we, users of Spider, benefit from it.
If compression of activities duration just add resistance then I imagine the mechanical equivalent as compressing a set of springs with different resistance and minimum length. The procedure would be like pushing the assembly from both ends.
About your book I will take a look on how to get it, believe it is available on amazon.
Thanks for taking your time,
Best Regards,
Rafael - 0Stephen Devaux ·"Thanks for taking your time"
My pleasure! It’s refreshing to chat with someone who obviously takes these issues seriously! I wish people working in industries where project time is not just measured in money, but also in human lives and suffering (e.g., pharma, health care, potable water wells) would take it as seriously!
Fraternally in project management,
Steve the Bajan - 0Vladimir Liberzon ·Hi Guys,
thank you, I enjoyed your discussion.
Rafael,
if you will look at my presentations I always insisted on setting single project success criterion estimating the cost of the time. If cost of time is estimated and known it is easy to model it by including hammock activity or activities in the project schedule.
Best Regards,
Vladimir - 0Rafael Davila ·
Vladimir,
I know about time cost and realize fixed cost will decrease with a reduction in time of the hammock activity contrary to crashing of an activity that will increase with reduction in time. But because it span across all project duration it will always be on the critical path therefore for purpose of determining where to crash is irrelevant, but it is still relevant for the determination of total crash cost, you are correct.
On the other hand the time cost related to individual activities I would consider it when determining crash cost of an activity this will vary from activity to activity and also cannot be overlooked.
I downloaded the MS Project add-on suggested by Stephen to study the walkthrough document of how to use DRAG to optimize a schedule on a sample project file. When your project starts to fall behind schedule you try to compress the schedule using other methods but crashing, these are managerial decisions no way you can pre-program into the CPM logic, and they are in essence changes in logic. It is after changes in logic no longer solve the issue when you have no other option than to crash some activities. Knowing how to crash, especially when your job is in problems is very valuable.
Many times I have given true chance to MS Project but always, very soon, it gets into my way, no matter how hard I tried, even MS Project 2010 is still a no-no. No way am I going to use it in my jobs to find in the middle it will let me down, I already know it will. I believe if Stephen techniques can be applied within Spider then why not give it a try, we already have DRAG embeded into Spider, better than an add-on.
Best regards,
Rafael - 0Vladimir Liberzon ·Rafael,
I meant something different.
It is necessary to define what is the cost of one day of project delay and the cost of one day of project finish acceleration.
Total cost shall be estimated taking this into consideration. And cost of crashing will increase or decrease this total cost. If crashing decreases this total cost it makes sense.
I did not suggest to estimate time cost for individual activity. For decision making you shall look at the total cost.
Best Regards,
Vladimir - 0Rafael Davila ·Vladimir,
Agree 100%. But you still need some individual activities cost of crashing. The following to me would be practical and acceptable, no need for pretentious non linear functions, just start with normal duration and crash duration. Targeting for larger than normal does not makes sense.
The above should simplify the math and the data entry.
Best Regards,
Rafael - 0Vladimir Liberzon ·Usually there are more than one way of crashing an activity. And usually crashing is not linear and continious. Adding additional resource you will shorten activity duration by certain amount of time. But this resource will not be used on one activity only, so it is much more complicated. But different options may be weighted by looking at the new total cost that includes the cost of time.
Best Regards,
Vladimir - 0Vladimir Liberzon ·Rafael,
I understand time-cost trade off on the project level and not on activity level.
Changing one activity duration may change the whole schedule and create new resource critical path. Besides the result depends on resource leveling priorities.
But in any case usually when you decide to implement some action like adding resources or making people to work on vacation day it will change not only one activity duration but also some others.
The result shall be analyzed on the project level and you can play "what if" games to find satisfactory solution.
When you decide to crash the schedule you think not about just shortening some activity duration but about something real - use additional resources, different technology, or increasing resource work time. Any measure is complex and will lead to many changes in your schedule.
An impact of your measures shall be estimated by rescheduling the whole project.
Best Regards,
Vladimir - 0Vladimir Liberzon ·In the new version of Spider Project materials that are required or consumed on activities can be shown the same way as renewable resource assignments - under activity lines.
In projects with large amount of materials it is much more convenient than looking in the columns.
Best Regards,
Vladimir - 0Rafael Davila ·Vladimir,
“When you decide to crash the schedule you think not about just shortening some activity duration but about something real - use additional resources, different technology, or increasing resource work time. Any measure is complex and will lead to many changes in your schedule.”
Yes I realized that in my simplified mechanical model the spring constant might represent resistance to compression (a cost issue) but the element of resources is missing, therefore compression of the schedule is not that simple even when disregarding the extra resources for the activities being crashed you still cannot disregard original resources. Anyway I was kind of uncomfortable with not considering the new resources in my wish to find a simplified procedure.
About material resources being shown in the Gantt chart this is quite convenient; in what version number it will be displayed? I just downloaded new version 10.01.69 but could not see them. Are these resource bars to have different color than regular resource bars? I was looking at the sample construction job, so you can make reference to it.
Materials resources I do not use much, but when needed there is no substitute for the modeling of these, still on my mind the modeling of space as a material resource.
Best regards,
Rafael - 0Rafael Davila ·Vladimir,
Thanks, the materials display is very welcomed and so will be the Materials Gantt as fields particular of material resources will be available on the table view.
I would not oppose if a Time Distribution Table View is also added for resources, materials and costs. I do not pretend the view to be a substitute to the tabular reports I can generate with my templates, it is a matter of speed while navigating on the computer screen. Perhaps something faster to setup at the expense of the versatility of the report generator. It would be great if by adjusting time scale it would automatically adjust periods. I am used so much to the slider to adjust timescale that whenever I use other software the lack of this simple functionality annoys me.
I was also thinking about scrolling vertical and horizontal icons bars. The vertical similar to MS Project, this concept can be applied to the horizontal bar. In this way if you have a reduced window you will still be able to reach all icons. I do not like menus to be customizable as every time I must use other person computer where the menus have been customized I get lost, I prefer standardization of menus.
I clocked the real time to download last upgrade, it took less than 2 minutes and install less than one, is this is an improvement on the server? This is how all enterprise and portfolio capable software should be. Sorry but good things you got to celebrate. Download and install takes two clicks of the mouse maybe you can set it up to auto-install to save one click.
Best regards,
Rafael - 0Vladimir Liberzon ·
Optimization +
We introduced next level of optimization that may improve resource-constrained scheduling results in complex cases.
We call it optimization+ but maybe we will decide to keep this and remove previous optimization option.
We decided to add this level after testing Spider algorithms on a set of sample projects specially created for testing levelling capabilities of PM software. Though Spider achieved the result that is much better than achieved by any other software we decided to make the difference even larger.
Pro - the scheduling results may be improved, Cons - takes more time.
By the way, in large and complex projects we suggest to optimize project schedule and then use Previous version support option that works very fast. In selected previous version the schedule is optimized.
From time to time the schedule shall be optimized to check if another order of works is preferable after implementing project changes.
Other news soon.
- 0Rafael Davila ·
Vladimir,A few moths ago I
Vladimir,
A few moths ago I read on the internet that optimization algorithms although all search for the optimum how close will depend on the particular schedule, some schedules will be better optimized by a certain algorithm while others by other algorithm. Perhaps the response is so complicated that only by running the different algorithms you will know which is best for your particular schedule.
If Optimization and Optimization + target the same group of schedules then is ok to remove the first, and keep the space for other algorithms that target other schedule groups.
The trick to select previous optimization after selecting optimization plus in order to speed it up is a good one, thanks for the advice.
I have noticed (I believe), that the use of certain functionalities such as Skills can also impose some extra burden on the resource leveling and therefore can slow the computations. Am I correct? I ask because most of my schedules are of about 600 activities, so small that the difference is not much.
Best regards,
Rafael
- 0Vladimir Liberzon ·
Rafael, if to use simple
Rafael,
if to use simple priorities that we call standard then you right - one algorithm is good for one type of schedules, another - for another type of schedules.
But optimization is the method that will produce good decision for all types. Even if not optimal but close to optimal.
Optimization+ in some cases may improve the schedule calculated with optimization method but is slower.
What can make resource levelling slower: skills, negative lags, start to finish links, material and cost levelling, hammocks with assigned fixed amounts of materials and costs. If resources are very scarce the levelling will be slower that in the schedules where resource shortage is relatively small.
In any case it is not hard to optimize the schedule and then run Previous version support.
Besides, Previous version support option makes the schedule stable that is very important.
- 0Vladimir Liberzon ·
Material Gantt
New version of Spider Project includes Material Gantt. It is still under development, we shall add properties and reports. But many people asked for this view and we decided to make them happy and then add all we plan to include in this view.
I wanted to add the picture but it does not work the same way as in old forum. I shall learn how to do it.
- 0Rafael Davila ·
Vladimir,So far so good, but
Vladimir,
So far so good, but as you said it is under development, to me it is not just about the Gantt but also about the view giving access to material usage reports and some extra fields not available in other views. Are we going to have some reporting capabilities here?
By the way, about "dynamic links" that can be toggled on and off by formula execution my anxiety is based on the fear it might be an imposible dream, I fear that as schedule logic varies with these dynamic links then resource leveling will vary and if you run several times the schedule computation in some ocassions might converge to the same solution while in other ocassions it might not. Not sure about how this is to be implemented but fear the possibility of never ending "do-loops" and how are you to mange the issue if there is such possibility.
I feel like a kid to which a box of candies is shown but not allowed to reach it.
Best regards,
Rafael
P.S. You shall click "Disable rich-text" paste your links and go back to rich text to continue with your reply.
- 0Vladimir Liberzon ·
Material Gantt
Rafael,
reports will be added soon, but most of them can be created now. Wait, improvements will be added soon.
I don't expect loops but expect that the choice may depend on selected levelling priority and may change during project execution depending on actual performance.
Best Regards,
Vladimir
- 0Rafael Davila ·
Vladimir, This new view will
Vladimir,
This new view will add a new dimension to reporting, I like the idea of keeping it at the Main Window instead of at the Project Window, it keeps the core easily accessible, excellent choice.
That you do not expect loops is good news, give me some peace of mind, Dynamic/Conditional links I believe will add additional advantage to Spider over the rest.
Best Regards,
Rafael
Log in or join to reply
ProjectControls.com is free to join. Members can post questions, reply, vote and follow discussions.
Existing members: your original username and password still work. Forgot password?