Sucessive Amplification Methods

When dealing with multiple criteria decision problems, the Decision Maker may be confronted with a large number of nondominated alternatives, which is excessive for the Decision Maker’s limited processing capacity, although there is no difficulty in choosing among a reduced number of solutions. A possible strategy for nonprescriptive decision-aid on this type of problems consists on gradually approaching the preferred alternative, without surpassing, at any stage, the information processing limits of the DM. The Successive Amplification Method uses fuzzy clustering techniques to define a limited number of macro-solutions, which represent holistically the decision set. At each stage, the DM selects the most promising macro-solutions, and the actual decision set is reduced accordingly. This corresponds to a progressive amplification of the preferred area of the attribute space, until a final decision is reached. In the process, the DM has only to decide among alternatives (real or virtual). The paper describes the method and presents an illustrative example.