Membership function elicitation in possibilistic programming problems

In this paper, we propose an elicitation method for membership functions in the setting of possibilistic programming. Each possibility distribution is defined by several possible domains with different possibility degrees. Ta king advantage of the properties of possibility and necessity measures, a n ested target table for the objective function and for each constraint is el icited from a decision maker. To aggregate all the goals, an additional nes ted target table is established. Through those nested target tables, member ship functions can be determined. It is emphasized that the information of nested target tables is much more important than that of membership values in the setting of possibilistic programming. The membership functions in po ssibilistic programming are not assumed to be cardinal but ordinal. In this sense, the decision making via possibilistic programming can be regarded a s a qualitative one. (C) 2000 Elsevier Science B.V. All rights reserved.