Abstract
A temporal decomposition approach is presented in this paper to solve a long term deterministic model for optimal operations of multiple reservoir systems. Through Lagrangian relaxation, the long term problem is decomposed into a number of smaller subproblems. Each of the subproblems can be solved efficiently using standard NLP codes. Coordination between subproblems is achieved in a Lagrangian term. Overall convergence is attained through an iterative process by updating the Lagrangian multipliers, A theoretical proof of global convergence is given assuming a concave objective function. The method has been applied to a nine-reservoir system in central China.