Applying Monte Carlo Analysis – Risk Analysis
Key requirements for a QSRA
You will need a robust and fully logically linked schedule where only genuine constraints remain
If the project is in progress, the schedule must be updated to reflect current status
You must have identified duration uncertainty against all activities using the three-point estimation duration for all activities
Risks are identified, documented and their impacts assessed and allocated to the schedule activity
A software application, such as the two examples given above, are needed with which to run the Monte Carlo analysis, which produces appropriate reports
Terminal float or other time contingencies should be removed from the schedule prior to running the QSRA
The five stages of schedule risk analysis
Stage one – Schedule quality check
The first step in running a QSRA is to confirm that the schedule is of the required quality. Some other technical steps can be made by QSRA software, for example, open ends, and type of logic used.
Other matters may be more subjective, such as the general robustness of the schedule and the surety that all project scope is included. Ideally the schedule should be checked by an independent party prior to running a QSRA.
Level of effort-type activities should be excluded from the analysis, as they do not represent true work.
Stage two – Duration uncertainty
A quick risk analysis can be performed by setting a narrow uncertainty range across all activities. The purpose of this is to help highlight key activity drivers in the schedule and to ensure that the project logic is robust.
For a full assessment of duration uncertainty, the project team will need to be consulted to assess the “minimum”, “maximum”, and “most likely” durations for each activity to feed into the risk analysis.
This could be done on an activity by activity basis, but this would be extremely onerous unless the data is captured as part of the duration calculation when developing the schedule.
So it is usual to categorise different types of work and apply confidence factors to each type of work. Distribution types that can be applied are discussed below.
Stage three – Monte Carlo analysis, the first run
After applying and the duration uncertainties, it is useful to run a “first pass” as this will assist in understanding the final results, in other words, how much of the final result is attributable to confidence in the current schedule, and how much due to the assessment of potential risk events.
An understanding of this will assist in assessing and prioritizing mitigation actions. For example, a disproportionately bad result at this stage may point to the requirement for more planning to further develop or mitigate the current schedule.
Stage four – Adding the risks into the analysis
Once the risks have been identified and catalogued in the risk register, those that have the potential to cause a delay are added into the schedule.
Durations, probabilities and distribution types are added to each risk.
The probability is the likelihood of the risk event occurring, and is expressed as a percentage.
The duration is the expression of the likely effect in terms of time, and maybe at 1. Value, or a minimum/most likely/maximum range of effects.
Stage five – Monte Carlo analysis, the second run
This stage assesses the effect of both duration uncertainty and the risk events.
An understanding of this will assist in assessing and prioritizing mitigation actions.
For example, an unwelcome or unexpected result at this stage may point to the requirement for more mitigating actions to be developed, which may include re-working the network schedule.
It will be necessary to iterate through the two processes, assessing mitigation options until you have a solution that best meets the needs of the project or programme.