Multiple Criteria Decision-Making Process
Since almost all real-world decision problems must be addressed on the basis of multi-dimensional approaches, the Multiple Criteria Decision-Making (MCDM) process has been developed long time ago. Pareto first introduced the efficiency concept in 1896 (Doumpos and Zopounidis, 2002). A feasible solution is efficient if and only if there is no other feasible solution which dominates it (Ringuest, 1992). Von Neumann and Morgenstern (1944) developed the utility theory as one of the major methodologies in modern MCDM. The preference structures of a decision maker are represented by multiple attribute utility functions. Charnes and Cooper (1961) extended the traditional mathematical programming theory to the goal programming. In recent decades, more and more user – friendly software has been developed based on advances in information technology and computer science.
Basically, MCDM provides a set of criteria aggregation methodologies that focus on decision maker’s preference structures, system values and judgment policy. Since the “optimal” solutions in the traditional mathematical programming usually do not exist in MCDM due to the potential conflicting nature of the multiple objects, MCDM could find an appropriate “compromise” solution that satisfies all of the decision maker’s policy. MCDM general procedure consists of (1) identifying decision objectives, all feasible alternatives and participants; (2) developing evaluation criteria that measure the performance of each alternative on decision objectives; (3) modeling criteria aggregation; and (4) providing meaningful recommendations. MCDM approaches include multi-objective mathematical programming (MMP), multiple attribute utility theory (MAUT), outranking relation theory (ORT), interactive methods and preference disaggregation analysis (PDA) (Vincke, 1992; Pardalos et al., 1995).