Stochastic SDDP Cut Sharing

Units:	-
Default Value:	0
Validation Rule:	In (0,1,2)
Description:	Type of cut sharing applied in the SDDP formulation

Stochastic SDDP Cut Sharing is an advanced option for the SDDP algorithm.

The Cut Sharing option determines how the Benders cuts, generated during the backward pass of the algorithm, are shared between nodes (full branches) of the scenario tree. It can take one of the following values:

None (value = 0)

Cuts are not shared between scenario tree nodes.

Single (value = 1)

Cuts are shared between nodes and the subject of the cut equations is a single future cost function decision variable.

Multiple (value = 2)

Cuts are shared between nodes and each shared equation defines the future cost with a unique decision variable.

Cut sharing can improve SDDP algorithm convergence.