A comparison of multiattribute decision-making techniques using an iterative procedure to derive a convergent criterion

The objectives of the dissertation are to: (1) deductively compare five commonly used multiattribute decision making techniques and clarify the meanings of weights and (2) empirically evaluate the effectiveness of the techniques. The five techniques are multiattribute utility theory (MAUT), weighting and rating (WR), the analytic hierarchy process (AHP), concordance analysis (CA), and computation of equivalent alternatives (CEA).; From the deductive comparison, we concluded (1) that the tradeoffs among attributes can be transformed from one technique to another, (2) that the weights in these techniques have different meanings and can be explicitly identified, (3) that a valid criterion can be derived internally from a neutral procedure for empirical comparisons, and (4) that the application of WR and CA is limited because of their embedded assumptions.; Three techniques (MAUT, AHP, and AHP{dollar}spprime{dollar}) were compared empirically. The AHP{dollar}spprime{dollar} technique is a modified version of the original AHP to conform to the mathematical meaning of weights. Two experiments were conducted. In the first experiment, two techniques (MAUT and AHP{dollar}spprime{dollar}) were compared based on an iterative procedure. In the iterative procedure, the subjects applied one technique first, then the other, and iterated each until a stopping rule was met. The judgments in each iteration were anchored by the judgments made in or transformed from the preceding two iterations. The final judgments were neutral to the two techniques and, therefore, were the criterion based on which the techniques were compared. In the second experiment, MAUT was compared with AHP based on the same iterative procedure.; The hypotheses for the two experiments were (1) that AHP{dollar}spprime{dollar} is more effective than MAUT because of the ease of use and the elicitation questions conforming to the mathematics in AHP{dollar}spprime{dollar} and (2) that MAUT is more effective than AHP because the elicitation questions of relative importance in AHP are ambiguous. The results did not support either hypothesis. MAUT was more effective than AHP{dollar}spprime{dollar} in the first experiment, but not significantly different from AHP in the second.