Part 80: Representations of Uncertainty
(Project Risk Management: Perform Quantitative Risk Analysis)

  • We don’t always know what the exact duration or cost of an activity will be, but we might have a range (a set of values)
  • Each value in the range may have a different probability of occurring
  • We represent the values in a probability distribution
  • There are many types of distributions, including triangular, normal, uniform, and discrete
  • We should select the type of distribution that best represents our data
  • We might have a different distribution for each risk
  • For example, the cost of installing a window will probably be $1000, but between $500 and $1500.  We have a range.  The actual cost might be any value between $500 and $1500.  We have a 50% chance that it will cost $1000, we have a 20% chance that it will cost $800 or $1200, and we have a 10% change that it will cost $500 or $1500.  We have other percentages for other values.  This can be represented by a distribution.
  • Sometimes two or more risks are related.  We call this correlation and should include it in the model. For example, if a truck carrying toxic chemicals has a crash (Risk 1), then there is a possibility that the environment will be polluted (Risk 2)