A New Method of Sensitivity Analysis for Multi-Criteria Decision Making in Water Resources

Multi-Criteria Decision Making (MCDM) techniques are widely used as practical tools to select superior projects/alternatives in water resources. However, parameter uncertainty (e.g., uncertainty in weights and performances) can significantly affects their performance in such a way that making a robust decision is too much difficult for a decision-maker. Thus, it is necessary to evaluate the effect of uncertainty on ranking of alternative decisions in order to reduce the possibility of failure in selecting a robust decision. In this study, we develop a novel probabilistic methodology for sensitivity analysis of a MCDM model to determine the robustness in decisions considering uncertainty in parameters. The proposed methodology which is developed based on Advanced First Order Second Moment (AFOSM) method estimates both performance and the probability of failure in ranking of alternatives. This provides additional information for decision makers to select a superior alternative with minimum risk of failure. This technique is applied to two case studies in water resources which previously evaluated using Weighted Sum Model (WSM), and the results are compared with those obtained from alternative sensitivity analysis methods, such as Monte Carlo based sensitivity analysis techniques. The results show that the proposed method improves the results of Multi-criteria analysis for water resources problem by providing additional information about the robustness and risk of failure in rank of alternatives when it is compared with other sensitivity analysis methods.