What is a decision problem?

Abstract

This paper presents a general framework about what is a decision problem. Our motivation is related to the fact that decision analysis and operational research are structured (as disciplines) around classes of methods, while instead we should first characterise the decision problems our clients present us. For this purpose, we define a decision problem as a set partitioning problem and we introduce a new framework, independent from any existing method, based upon primitives provided by (or elicited from) the client and four fundamental problem statements. We show that constructing the set to partition is a decision problem itself and that the types of primitives are finite. The result is that the number of archetypal decision problems are finite and so the archetypal decision support methods.