Stochastic FCF Scalar
Units: | - |
Default Value: | 1000000 |
Validation Rule: | Any Value |
Description: | Scalar for future cost function objective and constraint terms |
Detail: |
FCF Scalar is a scalar for the future cost function objective and constraint terms in the SDDP algorithm. Scaling affects the numerical stability of the linear programming problems. It is good practice to experiment with the FCF Scalar to find a value that provides the most stable convergence for the given problem.
As a rough guide, if the total objective function value of your simulation is in the order 1E+9 then an FCF Scalar of 1E+6 might work well. If the objective is in the order 1E+10 then values up to 1E+7 might be better. Note that the scalar does not need to be a perfect power of 10 e.g. you can set a value of 5E+6.
See also: