Deltek Acumen Risk provides an efficient way to account for activity duration uncertainty in your schedule. All projects come with some level of risk. It’s the reality that besets the art and science of scheduling. And even a schedule with very few identifiable risks still has a degree of uncertainty in the duration estimates.
Most schedules do not account for duration uncertainty. Instead they rely on single point deterministic activity duration estimates that assume durations are known with a large degree of certainty. The reality is that durations are not certain, and are better modeled using a probability distribution. Assigning each activity a normal distribution or a slightly skewed normal distribution (beta distribution), in most situations, is tedious.
Example of slightly skewed beta distribution
A triangular probability distribution from a three-point estimate is more feasible. The triangular estimate identifies the pessimistic, most likely, and optimistic duration estimate for each schedule activity. This provides a significant improvement over the single point deterministic duration estimate (that which was derived by the critical path analysis).
The three-point estimate is a noble and achievable goal, but may quickly become logistically laborious on large projects. Attaining three-point estimates from team members and subject matter experts becomes an unwieldy effort on any sizeable project schedule. Deltek Acumen software, however, provides a handy way to make a three-point probability estimate of each activity. These three-point probability duration estimates are combined to find the schedule aggregate.
This article demonstrates activity duration uncertainty analysis using triangular probability distributions and Deltek Acumen.
We have in Figure 1 our demonstration project.
Figure 1
This is a simple piping installation project that consists mainly of three activities that have Finish to Start (FS) relationships. Let’s assume there are no known risk events. As mentioned, we still have to consider duration uncertainty. Single point deterministic duration estimates say the project concludes on June 14th, 2019. We want to consider triangular probability distributions of our schedule activities by implementing Deltek Acumen Risk.
Our goal is a triangular type duration uncertainty analysis to better predict the reality of activity duration schedule variance. We want to demonstrate that Deltek Acumen is an effective means for considering activity triangular probability distributions of duration estimates.
Let’s proceed in Deltek Acumen Risk and select to import a Primavera P6 XER file, Figure 2.
Figure 2
When the popup menu appears highlight our project, construction-01, and click OK, Figure 3.
Figure 3
Now we need to select Import all projects to actually import the schedule, Figure 4.
Figure 4
Great! Now, as displayed in Figure 5, our project is fully imported into Acumen.
Figure 5
To run a risk analysis we must first save our schedule in a workbook. In the upper left of the screen click the Acumen icon, save as, and Deltek Acumen Workbook, Figure 6.
Figure 6
Continue and click the tip S3 // Risk tab, Figure 7.
Figure 7
Next, click Run Risk Analysis, Figure 8.
Figure 8
Deltek Acumen Risk then runs a Monte Carlo simulation on the project. In Figure 9 we have our risk analysis results.
Figure 9
Note that the Duration Uncertainty settings, Figure 9, are turned off. No probability distributions are assigned. This risk assessment is for a single point deterministic analysis. On the Cumulative Frequency chart all iterations resulted in a completion date of June 14th, 2019. Without duration uncertainty, the Monte Carlo analysis had no duration ranges to work with so everything remained as-is.
Now we want to assign our schedule triangular probability distributions. First click the uncertainty icon, Figure 10.
Figure 10
Set the template size to 7, Figure 11, so we have seven categories for our triangular probability distribution.
Figure 11
Second, Figure 12, set the duration uncertainty to very conservative.
Figure 12
This means that the duration estimates provided by our team members and subject matter experts were conservative. We would therefore expect these estimates to hold true or better.
Look at the duration uncertainty values for the remove damaged piping activity, Figure 12. This activity is scheduled to take two days. However, note that the minimum duration at a 50% multiplication factor (MF) corresponds to 1-day, most likely duration at a 100% MF comes to 2-days, and the maximum duration at a 100% MF is 2-days. The fact that we have a minimum duration of 1-day at 50% MF, means that our triangular probability distribution is allowing for the possibility that this activity finishes early, which is good news.
Click the lower half of the run risk analysis icon and choose the uncertainty only (No risk events), Figure 13.
Figure 13
We are only considering duration uncertainty in our analysis. We assume that our project has no apparent risk events. Run the risk analysis, Figure 14.
Figure 14
Our results are displayed in Figure 15.
Figure 15
This time our iteration distributed among several dates. Observe the metric table below the graph. The first value says Deterministic 100%, 6/14/2019. This means that all iterations completed at least by the expected project completion date. And we had numerous iterations that completed earlier.
This is good news. Because we told Acumen Risk our duration uncertainty was very conservative it skewed the triangular probability distribution duration estimates, accordingly. Our very conservative triangular probability distribution meant our activities would meet their estimates or better. This includes the worst case that fell on June 14th.
In Figure 16 we set the duration uncertainty to very aggressive.
Figure 16
This means that our duration estimates were risky; it is not likely that our duration estimates will be met. So we would expect duration estimates to barely meet estimates or worse. Again, review the duration uncertainty values for the remove damaged piping activity, Figure 16. The minimum duration at a 100% MF corresponds to 2-days, most likely duration at 100% is 2-days, and maximum duration at 150% MF is 3-days. The fact that we have a maximum duration at 150% MF correlating to 3-days when the activity is scheduled to take 2-days is not good. The triangular probability distribution is considering the possibility that this activity finishes late, which is a negative.
The results of this Monte Carlo analysis are in Figure 17.
Figure 17
This time the first value is Deterministic 0%, 6/14/2019. This means that none of our iterations met the June 14th completion date, which is a problem. The worst case iteration was June 20th. Because we told Acumen Risk our duration uncertainty was very aggressive it skewed the triangular probability distribution to the right. Our very aggressive duration uncertainty setting meant that it was not likely that our duration estimates would hold true. They didn’t hold true and worse.
Summary
Deltek Acumen Risk provides an efficient way to account for activity duration uncertainty in scheduling. Schedulers can choose to turn off duration uncertainty for a single point deterministic analysis of the schedule.
Acumen Risk also provides the quick functionality to apply a triangular probability distribution to the entire schedule or individual activities. Acumen Risk computes a triangular distribution composed of minimum, most likely, and maximum duration estimates.
This triangular distribution is also skewed either left or right depending on the duration uncertainty setting, which can range from extremely conservative to extremely aggressive depending on your Uncertainty settings. So, the subject matters experts only need to provide one duration estimate for each activity, and also how conservative their duration estimate is. Acumen Risk takes this information and generates triangular probability distributions skewed left or right, accordingly. The results include the percent of meeting the deterministic estimate.
Additionally, Deltek Acumen Risk provides the best case scenario at zero percent probability and worst case scenario at 100 percent probability. Thus, Acumen Risk provides a quick and effective activity duration uncertainty risk analysis. Most projects though have associated risk events, which in Acumen is fine. In addition to duration uncertainty, Acumen supports both the definition and mitigation of risk events.