Fuzzy partitioning and its application to reservoir operation problem (a multistage approach using Markov chain)

Abstract

In this paper, we investigated mathematical models on reservoir operation problem and provided a fuzzy model based on Markov chain time series for Karkheh dam in Iran. Based on fuzzy partitioning of monthly streamflows, calculated by time series, a Markovian forecasting model was developed. A deterministic and a fuzzy partitioning stochastic dynamic programming model were formulated for the problem. The goals were formulated with weighting priorities and the optimal reservoir operation was determined. Using historical data, the performance of the dam was simulated. Results of this simulation clearly showed that the fuzzy partitioning proposed stochastic model outperforms the deterministic model.