Decision Analysis: Decision Trees, Expected Value & Monte Carlo

What you'll build

What you will be able to do

Each lab ends with a concrete result that you build yourself and the checker confirms, cell by cell.

• Compare three ways to build the K2 robot under three demand states and see that four classic decision criteria give three different answers.
• Roll back a two-stage tree for the K2 line: average at chance nodes, choose at decision nodes, and find the value of building.
• Price a perfect demand forecast for the K2 robot and check that EVPI equals the smallest expected opportunity loss.
• Update the demand probabilities with Bayes' rule for both survey results, find the best decision after each result and price the survey.
• Value the three K2 options with an exponential utility function and see that a risk-averse company picks a different option than expected value does.
• Turn 60 given random numbers into 30 possible futures for the small plant, then read the expected NPV, the chance of a loss and the shape of the distribution.
• Swing each driver from its P10 to its P90 estimate, rank the drivers by NPV swing and share of variance, and draw the tornado.
• Value the K3 robot three ways – invest today, wait a year, invest with an expansion option – and see that flexibility has a price tag.