Decision under risk: incomplete information and multiple objectives

Summary The expected value criterion can be extended to the case of incomplete information concerning the value function and probabilities. There are two questions concerning decisions under incomplete information: – How can the incomplete information be modeled? – How can preference statements be obtained from incomplete information? Preference statements can often be derived using simple linear programming (LP) approaches. Risk analysis is an important method for decisions under incomplete information concerning the value function. Sensitivity analysis is another possibility for supporting decisions under incomplete information. In the case of multiple objectives, preferences can be modeled by multiattribute utility functions. The simplest form is the additive utility function, which is applicable if the requirement of the additive utility dependence of the attributes is fulfilled. The condition for the mutual utility dependence is less restrictive. It implies a multiplicative utility function.