Abstract
Results from testing of the sampling stochastic dynamic programming (SSDP) methodology on long term hydropower schedul-ing problems compared with stochastic dynamic programming (SDP) was presented. Previous studies have shown promising results using SSDP on similar problems. A main difference between the methods is how uncertainty is represented. The SSDP method has been implemented and tested on aggregated regions from the Norwegian hydropower system. We have in this study not found one method to consistently outperform the other.