This paper presents a two-phase artificial neural network approach for find
ing optimal scheduling of interconnected hydropower plants. In this approac
h the objective is to maximize hydropower generation and to satisfy the irr
igation requirement as far as possible. A case study considering scheduling
of Bhakra-Beas interconnected reservoir system Is also presented in this p
aper. The proposed technique is compared with augmented penalty function me
thodology and demonstrates the potential of achieving better results.