A decision support system for discrete multiple-criteria problems: certainty, uncertainty, and hierarchical

Abstract

In this paper, we present a Decision Support System (DSS) framework for solving discrete Multiple Criteria Making (MCDM) problems. This is the first description of an integrated fourth generation MCDM DSS. Different Decision Maker (DMs) think, learn, behave, and make decisions differently. Hence, one type of questioning/answering, procedure, or assumptions about the DM's behavior may not sufficiently support DMs to learn about their preferences and make appropriate decisions. To provide a variety of flexible and applicable tools for the DM, a set of approaches are described that use different types and numbers of questions, different procedures to solve the problem, and different assumptions regarding the behavior of the DM. Emphasis is placed on providing simple mechanisms within the DSS that the DM can understand and can manipulate the DSS to obtain the desirable outcomes. These include mechanisms for the if, then type of questions. The DSS MCDM consists of six different modules: a) Data entry, assessment, and statistics; b) Screening methods, using various types of efficiency definitions and partial information; c) Complete ranking methods, using simple, additive, and concave/convex quadratic forms; d) interactive methods for finding the best alternative, including quasiconcave and quasiconvex, and any nonlinear forms; e) Extensions to problems under uncertainty for payoff matrices; and f) Extensions to hierarchical multiple criteria problems, including decomposition, assessment, and aggregation for large numbers of criteria. Each category provides several subroutines/procedures to accomplish the given task. Several examples, features of the developed software package, and some experiments with it are discussed.