Optimal operation model for reservoirs using mixed independent and Markovian inflow assumption

Stochastic Dynamic Programming (SDP) is a powerful optimization technique frequently used in planning optimal reservoir management. This method suffers from the problem of dimensionality. In order to reduce the problem of dimensionality an innovative method namely Mixed Independent and Markovian Inflow Stochastic Dynamic Programming (MIM-SDP) has been developed. In MIM-SDP, if the inflow in certain periods has statistically insignificant serial correlation, inflow in those periods is treated as independent and for the other periods inflow is treated as Markovian first order. MIM-SDP is applied to a real reservoir system and results obtained were compared to results obtained from SDP. The results compare well with the SDP results and indicate the potential use of MIM-SDP.