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 Hammock

12 replies [Last post]
Rafael Davila
User offline. Last seen 1 hour 43 min ago. Offline
Joined: 1 Mar 2004
Posts: 4620

I am trying to automatically create several hammocks whose start will be defined by early use of a given resource and its finish will be determined by the late finish of the resourceuse by its predecessors SS/FF. The resource FM1 TO FM10 will be allocated using skills and will vary when changes in duration of other activities occur. 

I want to be able to model such variations automatically in a Monte Carlo simulation as to explore the cost probabilities. If using 9 Freezing Machines the deterministic cost will be less but the float availability will also be less, I suspect the option for 10 Freezing Machines will be a better choice.

FROM RESOURCE GANTT

FMH02 photo FMHammocks02_zpsi2m3jgec.jpg

FROM ACTIVITY GANTT

FMH01 photo FMHammocks01_zpsenuv5mcj.jpg

http://www.mediafire.com/download/xa4lqyo4ltxxt3d/RD601.009.sprj
http://www.mediafire.com/download/7tx1v31locf4ca6/RD601.010.sprj

Any ideas?

Replies

Rafael Davila
User offline. Last seen 1 hour 43 min ago. Offline
Joined: 1 Mar 2004
Posts: 4620

In order to compare the results if using different probability distribution curves, lognormal vs. triangular a model for each scenario was issued.  The following diagram summarizes the results.

The cumulative distributions to the left represent models using lognormal distribution while those to the right represent models using triangular distribution.

Using 9 FM [Freezing Machines] is the best deterministic option while using 10 FM is the best among lognormal options set  and the best among triangular option sets. Consider planning for 10 FM as the "best" option.

SA01 photo SA01_zpsclgwtwjz.jpg

Rafael Davila
User offline. Last seen 1 hour 43 min ago. Offline
Joined: 1 Mar 2004
Posts: 4620

Thanks Vladimir, so easy no instruction needed, a single click of the mouse for every version you want to add.

Best Regards,

Rafael

 

 photo MMM_zpsdrgrtxzy.jpg

Sorry Rafael,

I am used to the name S-curve applied to cumulative cost.

By the way Spider Project can also show probability curves of any number of versions of the same project simultaneously.

It looks like this:

 photo Probabilities_zpslwmp5wg5.jpg

Rafael Davila
User offline. Last seen 1 hour 43 min ago. Offline
Joined: 1 Mar 2004
Posts: 4620

Vladimir,

Keep in mind these are probabilities S Curves from Monte Carlo run with ASAP as a general rule.  And yes it will include about 6 scenarios.  I have always been curious about comparing success probabilities curves of different scenarios and Spider gives me all the tools I need as-is. I never expected for it to be at a single click of the mouse as if using report templates for deterministic S-Curves, so easy I rarely use templates for such simple task.

It was on the modeling of the "resource hammock" in a dynamic way for Monte Carlo as to automatically adjust staffing and their costs that I was stuck. I do not believe this is possible with software like P6 or MSP. To me this makes the exercise twice as interesting, not to mention the need to model spatial resources something for sure neither P6 or MSP can do.

In the exercise statement I said - If using 9 Freezing Machines the deterministic cost will be less but the float availability will also be less, I suspect the option for 10 Freezing Machines will be a better choice.

Well it is starting to look like option for 10 Freezing Machines will be a better choice.

Best Regards,

Rafael

Any two versions (scenarios) early and late S-curves may be shown in standard Spider report, but not many scenarios simultaneously.

Spider exports to text and directly to Excel.

Rafael Davila
User offline. Last seen 1 hour 43 min ago. Offline
Joined: 1 Mar 2004
Posts: 4620

Eventually all scenarios S Curves will be plotted, the following only includes 2 scenarios. Wee will be able to see how off-set the curves are as well as the intersection points, if any.

SCurvesW00 photo SCurves_zpsu16ax0ym.jpg

Rafael Davila
User offline. Last seen 1 hour 43 min ago. Offline
Joined: 1 Mar 2004
Posts: 4620

It is not difficult to export the distribution data and get descriptive statistics. From the data you shall be able to plot on same chart the S curves of various schedule versions for comparison purposes. The display of the charts is great but the ability to export the data opens up the possibilities for further analysis.

DSFM10 photo DSFM10_zps4zq17zko.jpg

Rafael Davila
User offline. Last seen 1 hour 43 min ago. Offline
Joined: 1 Mar 2004
Posts: 4620

Vladimir, deterministic models as well as probabilistic models are coming out as expected but a few more are needed.  For the moment I am somewhat busy and will take me several days to create the others even if using current models as templates.

Several Models photo Several Models_zpsphbjovg7.jpg

In addition P85 value can be another metric of concern.

I will expand on my reasons for:

  • Using the "resource hammock" as to model resource staffing that will include idle time cost.
  • Using Monte Carlo.
  • Using spatial resource model instead of fixed logic that wont necessarily do it in Monte Carlo run.

This exercise is based on the classic scheduling problem for the Westerscheldetunnel.

  • https://feb.kuleuven.be/rebel/jaargangen/2001-2010/2007/TEM%202007-3/TEM_07-3-07_Vanhoucke.pdf 

I will cross reference this discussion into Lets Challenge Spider in due time.

Best Regards,

Rafael

Rafael Davila
User offline. Last seen 1 hour 43 min ago. Offline
Joined: 1 Mar 2004
Posts: 4620

Vladimir,

WOW your solution works beautifully, so good that I found an error in one of my manually defined hammock for FM10. I used a minor adjustment to the following formula for a deterministic test schedule.

DurPlan = If (R="";DurPlan;0)
Execute in Activity Gantt after scheduling N, after cost calculation Y [for Monte Carlo] N [for Deterministic], in performance archive N

For a re-schedule/update of the deterministic schedule the formula must be executed manually or by script before schedule calculation, something I believe can be a formula option.

Latter on will work on the Monte Carlo run, but the test is a proof it will work as good.

Thanks a lot

Rafael

Sorry for the duplication - one of previous two posts may be deleted.

Rafael, I enjoyed your task. It deserves to be put in "Let's challenge Spider" discussion. It is pity that this discussion does not continue.

It can be done using formulas and certain procedures.

I understood that you wanted to create hammocks for modeling costs that depend on the elapsed duration of Freezing Machines use.

1) Create project activities with codes FM01, FM02, ... with zero duration and of duration type.

2) Set hour costs for these activities,

3) Set the same codes FM01, FM02, ... for your Freezing Machines (resources),

3) Create User Field with code R (or any other) and enter R (or other text) in the cells of this field at activities FMo1, FM02, etc. rows.

Looks like this:

 photo Project formula_zpsp6j2y955.jpg

Now set formulas like this:

 photo Formulas_zpsmvq44smx.jpg

 

Spider Project will apply first two formulas after calculating project schedule and will set starts and finishes of "resource" activities (FM01, FM02, etc.) equal to starts and finishes of corresponding resource usage.

Then it calculates project cost and then makes "resource activities" duration equal to zero.

This will be done in every Monte Carlo iteration and you will get all required probability distributions.

Formulas order matters.

And please download the latest version where we fixed minor error in Monte Carlo simulation.

It can be done using formulas and certain procedures.

I understood that you wanted to create hammocks for modeling costs that depend on the elapsed duration of Freezing Machines use.

1) Create project activities with codes FM01, FM02, ... with zero duration and of duration type.

2) Set hour costs for these activities,

3) Set the same codes FM01, FM02, ... for your Freezing Machines (resources),

3) Create User Field with code R (or any other) and enter R (or other text) in the cells of this field at activities FMo1, FM02, etc. rows.

Looks like this:

 photo Project formula_zpsp6j2y955.jpg

Now set formulas like this:

 photo Formulas_zpsmvq44smx.jpg

 

Spider Project will apply first two formulas after calculating project schedule and will set starts and finishes of "resource" activities (FM01, FM02, etc.) equal to starts and finishes of corresponding resource usage.

Then it calculates project cost and then makes "resource activities" duration equal to zero.

This will be done in every Monte Carlo iteration and you will get all required probability distributions.

Formulas order matters.

And please download the latest version where we fixed minor error in Monte Carlo simulation.