Evaluation of Fuzzy Rule Generation Methods for Reservoir Operation

Fuzzy logic (FL) as a soft computing method based on human inference is widely used for water resources issues such as reservoir operation. However, common to any FL modeling approach, the definition of membership functions (MF) and the inference rules (IR) are a difficult and time consuming practice that construct the objective of this paper. Self organizing feature maps (SOFM) method is applied to create MF. Also Fuzzy associative maps (FAM) and Bootstrap fuzzy relevance test (BFRT) are applied to generate IRs. Zayandeh-rud dam in Esfahan Province, Iran, was taken as the case study. The results showed that the suggested classification by SOFM to define MF is quite acceptable. But, the resulted rules from FAM and BFRT need more evaluation and screening before application. Nevertheless, the applied methods play a significant role in reducing the required time to set up the final fuzzy model.