Development of a stochastic optimization for multireservoir scheduling

Seasonal reservoir scheduling for multireservoir hydrothermal power systems is a practical, nonlinear, stochastic problem of high dimension. Extension of an existing deterministic algorithm to handle stochastic aspects is desirable. A linear feedback rule, but with chance constraints on lower limits only, gives some promising results. Exact solution by stochastic dynamic programming is possible only for a one-reservoir model. This result is compared to the deterministic and linear feedback solutions.