Tons of nodes in the Policy Graph, SDDP.jl package

So the “simulation” column has values on O(10^10). You can think of the mean of values in this column as an upper bound. (It’s not exactly, but near enough).

The “bound” column is a true lower bound. it has values O(10^8).

That means you are very far away from optimality, and that you need many, many more iterations.