Decision problems under uncertainty based on entropy functionals

This essay intends to define the role of entropy, in particular, the role of the maximum entropy criterion with respect to decision analysis and information economics. By considering the average opportunity loss interpretation, the basic hypothesis for Shannon's derivation can be derived from properties of decision problems. Using the representation Bayes Boundary it is possible to show that selecting a single probability from a set by the Maximum Entropy Criterion corresponds to a minimax criterion for decision-making. Since problems of randomly accessing and storing information as well as communicating information can often be stated in terms of coding problems, this result might be used to develop strategies for minimizing retrieval time or communication costs.