Preference Metric: Utility Functions
Monotonic Preference
- Increasing: [x1>x2] Û [u(x1)>u(x2)]
- Decreasing:[x1>x2] Û [u(x1)<u(x2)]
Attitude towards Risk
- Risk averse:
- Risk prone:
- Risk neutral:
Exponential Formulation
- Constant risk premium
- Works well in whole range
Other Formulations
- Polynomial formulation
- Logarithmic formulation
- Compound formulation
- Describes the logic between attribute bounds
- Best Case Utility: Ubest=1
- Worst Case Utility: Uworst=0
- Middle points: non-linear curve fit