Doasa is a version of the stochastic dual dynamic programming technique applied to the solution of hydro-thermal scheduling problems. The model can be used to formulate a policy of releasing water from reservoirs for electricity generation while satisfying demand over a fixed time horizon and minimizing the expected fuel cost of thermal generation. But the model will not necessarily satisfy demand in all scenarios. Given an appropriate penalty cost, some load may be shed.
The Doasa model has been developed by Stochastic Optimization Limited (SOL). The particular implementation SOL has made available to the Authority is described in the Doasa documentation.
The Doasa algorithm divides a year into 52 weekly stages and is restricted to the New Zealand electricity system. All input data are deterministic except for weekly inflows that are assumed to be stagewise independent. However, a technique known as Dependent Inflow Adjustment (DIA) may be applied to the empirical inflow distribution to make it share some properties believed to be present in the true inflow distribution.
Doasa is still being prepared for final release. Please email email@example.com if you wish to obtain a beta release.
Last updated: 28th November 2013