Developing optimal release policies of multipurpose reservoirs is very complex, especially for reservoirs within a stochastic environment. Existing techniques are limited in their ability to represent risks associated with deciding a release policy. The risk aspect of the decisions affects the design and operation of reservoirs. A decision-making model is presented that is capable of replicating the manner in which risks associated with reservoir release decisions are perceived, interpreted and compared by a decision-maker. The model is based on Neural Network (NN) theory. This decision-making model can be used with a Stochastic Dynamic Programming (SDP) approach to produce a NN-SDP model. The resulting integrated model allows the attitudes towards risk of a decision-maker to be considered explicitly in defining the optimal release policy. Clear differences in the policies generated from the basic SDP and the NN-SDP models are observed when examining the operation of Aswan High Dam (AHD). The NN-SDP model yields policies that are more reliable and resilient and less vulnerable than those obtained using the SDP model.
A stochastic dynamic programming framework for weed control decision making: an application to Avena fatua L.