Passive and Active Compensability Multicriteria ANalysis

Determining intercriteria relations is an important issue in MCDA. In order to handle this problem, this paper presents a new approach to the concept of compensation in multicriteria analysis, named PACMAN (Passive and Active Compensability Multicriteria ANalysis). The notion of compensability introduced here is decision maker oriented, relying only on information provided by him. In PACMAN compensability is analysed by taking into consideration two criteria at a time and distinguishing the compensating criterion from the compensated one. Separating active and passive effects of compensation allows one to point out a possible asymmetry of the notion of compensability. A valued binary relation of compensated preference is introduced. Compensated preference can be used for a modelization of the preference structure.