Task 4: Impact Analysis


With this task, we want to find out what are the benefits for the using the stochastic model with recourse.

There are two quality metrics: the VVS and the EVPI

Expected Value of Perfect Information (EVPI)

  • Individual determination of the best possible solution for each scenario
  • Difference between the weighted average of the best possible solution for each scenario and the objective function value of the stochastic solution
  • Interpretation: Maximum value that a decision-maker would be willing to pay in order to obtain complete information about the realisation of the uncertainty

Value of the Stochastic Solution (VSS)

  • Comparison of the deterministic and stochastic solution
  • Determination of the expected costs of the deterministic solution across all scenarios
  • VSS: Difference between the objective function value of the stochastic solution and the weighted costs of the deterministic solution
  • VSS=0 means that there is no advantage in taking uncertainty into account
  • VSS>0 shows that the decisions can be adjusted in anticipation of the uncertainty so that the expected costs fall
Sub-Task 1:  Implement a model that takes the result from the average run and set all investment decisions (y0) to the result vector, and then just optimises the rest of the core model.

Sub-Task 2:  Calculate the EVPI and VVS. (There is a jupyter cell that takes the values and creates an excel sheet, you can do the calculation in that excel)