Project simulation 2

[rajawaqar]

Hello All.

I am in Project Controls field. Scheduling and cost control reporting is my main job description. I am currently using PE&C for scheduling and VBA for reporting in Excel.
I am wondering if some body will help me giving some know how of as to how can I use simulation techniques (e.g. monte carlo) and computer modelling techniques (in Excel) for project analysis. I have knowledge of both of these techniques but it is limited to what I have learned during my Masters in Project Management in School.
Any help or guidance will be highly appreciated.


Raja Ahmed B.E. M.Eng.
Alberta, Canada.

 

[IRstuff]

Generally, you buy a tool that does that for you. @Risk is one such tool that uses Monte Carlo techniques to determine probable performance.

Essentially, you apply random numbers to all your durations to see what pops out.

But a good critical path analysis should do that for you as well.

TTFN

 

[rajawaqar]

Thanks for the quick reply.

I am familiar with @Risk as well. So what I get from your reply is that I can simulate the project durations, and that will in turn tell me whats gonna be the end date.

How can a critical path analysis will help? IS it that I need to load schedule with simulated durations and see if I have a critical path or not?

Thanks in advance.

Raja Ahmed B.E. M.Eng.
Alberta, Canada.

 

[IRstuff]

No, your critical path(s) are those that must be completed in serial fashion to get to the end of the project. That's the longest pole in the tent. Monte Carlo simply tells you, based on the parameters you give it for each task, what the most likely completion date is for the various critical and non-critical paths.

In Microsoft Project, you'd enter each task with its scheduled duration, provide the precedence links, and so on. The critical path is the one that delays the program on a day for day basis.

A program like @Risk will ask you for additional parameters per task, e.g., best-case/worst-case durations and some sort of probability function.

My earlier point is that if you have a handle on the critical path and have some understanding of the probability of success of each estimated task, which you essentially need for @Risk, you can do a crude estimate of the probable schedule variance by hand by essentially RSS'ing all the probable variances in critical path.

Something like @Risk will do that for all paths, so you might find that a lesser path has a higher probable impact to your schedule.



TTFN

 

[Akhibi]

I have not used monte carlo for schedule but I have developed many simulators for cost analysis using excel.

From what you said initially, I believe you have a good understanding of vba. If so, you can create a model of the project in excel and link all the schedules relationships using relational formulas. eg if pile capping has to start 4days after piling, you can say capping start date is (piling finish + 4).

You can also generate random numbers by activating the excel add-in "Analysis ToolPak" and applying the function "=RANDBETWEEN(A,B)" to generate random numbers between A&B.

Like I said, I have not used this approach for schedules but it works very well for reliability and cost simulations.

Greg Akhibi
Clearwaters Consulting Ltd

 

[rajawaqar]

Thank you guys very much. I am gonna try the hints provided and will simulate the end dates of my project. I hope it will make sense at some point in time. I just wanted to apply this technique and see what results I get.
Anyways thanks again and have a good day.

Raja
AB, Canada.

 

[dbuzz]

At a more basic level, MS Project's PERT feature allows you to do a three point estimate of the schedule. You enter optimistic, likely and pessimistic durations for each task and MS Project calculates the expected duration.

@Risk's MS Project module allows you to do a Monte Carlo simulation of a schedule (

http://www.palisade.com.au/riskproject/).

Remember though that these are just tools -- garbage in equals garbage out!

 

[zeusfaber]

From experience, it pays to be fairly clear in these matters about the difference between Estimating Error and Risk (with the first being a manifestation of the fact that it never takes exactly the same amount of time to do the same task, and the other recognising a sudden change to the work that has to be done in response to an event which may or may not materialise).

Some of the earlier simulation methodologies tried to treat both of these sources of variation in the same way, often with totally disreputable outcomes.

A.