Catching a plane example and EVIU

Revision as of 21:41, 21 December 2022 by Lchrisman (talk | contribs) (Changed "Free 101" to "Free Edition")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

It's a familiar decision -- what time should you leave home to catch an early morning flight? You'd prefer to enjoy your morning snooze for as long as possible. But you don't want to miss your plane!

Description: This simple decision-analysis model shows how to figure out the best time to leave -- to balance the risks of wasting time waiting at the airport gate when you could have been snoozing against the risk of missing the plane. It shows how to model the uncertainty about how long you'll need to go from home to the gate. It also illustrates an innovative metric, the Expected Value of Including Uncertainty: The EVIU is the value you get from explicitly considering the uncertainty instead of simple ignoring it and assuming that all uncertainty quantities are fixed at their median value.

The influence diagram

Influence diagram for plane catching problem.png

If you ignore uncertainty, your loss (time wasted -- red line below) is large if you leave home less than 80 minutes before scheduled plane departure. And is zero at exactly 80 minutes. But, if you consider the uncertainty in how long it will take to drive to the airport, go from parking to the gate, and the time at the gate, the expected loss is a smooth curve as a function of when you leave home (the green line). The minimum expected loss is at around 110 minutes. So that's when you should leave home. If you allow just 80 minutes, the probability of missing the plane is about 50%.

Graph comparing time wasted ignoring vs considering uncertainty.png

Download the model: Plane catching with UI 2020.ANA When you click to download, the model should start up in Analytica if you have it installed. If not, download Free Analytica Edition installer so that you can run and edit the model.


Video: Catching a plane example and EVIU (49 minutes)


Author: Max Henrion, PhD.

For more:

  • See this engaging video of Max Henrion in conversation with Alex at Risk Awareness Week in October 2020. Max introduces the problem, builds the model in Analytica, and explains the concept of the EVIU. Max also cites Archibishop Ussher of Ireland who calculated the exact day and time of the Creation of the World, and Socrates, who famously said "If I am thought wise, it is only because I know that I know nothing” (Phaedrus, sec 235) and annoyed his fellow Athenians so much that they condemned him to death.
  • Here are Max's slides: Media:Henrion Plane catching at RAW 2020.pdf
  • For more details on the EVIU: chapter 12 "The Value of Knowing How Little You Know" of Uncertainty: A Guide to Dealing with Uncertainty in Quantitative Risk and Policy Analysis, M. Granger Morgan and Max Henrion, Cambridge University Press: New York, 1990. ISBN 0-521-42744-4.
  • "The Value of Knowing How Little You Know", Max Henrion, PhD Dissertation, Carnegie Mellon University, 1982.

Keywords

Decision theory, decision analysis, uncertainty, Monte Carlo simulation, value of information, VOI, EVPI, EVIU, loss function, ignore uncertainty.

Comments


You are not allowed to post comments.