On Methods of Decision-Making under Interval Probabilities

Decision-making problems are generally classified into three types traditionally: certain one, uncertain one and risky one according to the decision-makers' information of states which they are facing. However there are some decision-making problems lying out of the classification scheme. Interval probabilities were presented to describe these situations and were developed to be a general theory to describe uncertainties and completeness in decision-makings. And to convert interval probabilities to point probabilities is generally accepted way to solve decision makings under interval probabilities. Our focus is mainly on how to convert interval probabilities to point probabilities. Three methods, computer simulation method, linear analytical method and maximum entropy method are put forward to convert interval probabilities to point probabilities, which make decision-makings under interval probabilities can be solved by the classical EMV principle. And their advantages and disadvantages are discussed by the end.