2 METHODS FOR LARGE-SCALE NONLINEAR OPTIMIZATION AND THEIR COMPARISONON A CASE-STUDY OF HYDROPOWER OPTIMIZATION

This paper presents two methods for the optimization of structured lar ge-scale problems: a decomposition method of dual type for nonlinear p roblems and a sequential quadratic programming based method. Practical details of application of the methods to the case study problem of th e hydropower system of an African river are then given. Comparison of results is presented, indicating that both methods are useful and effi cient, having however different features from a practical point of vie w. General remarks concerning the practical differences between a deco mposition-based method and a method exploiting the problem structure w ithin the framework of general purpose optimization routines are final ly presented.