Open MADNANTARRAR opened 2 weeks ago
Can you tell us what is the tolerance for DDP algorithm convergence you have used? There are likely multiple near-optimal solutions for this problem (i.e. small differences in cost resulting from stage 3 operations, which has lowest impact on objective function). It possible that model may be terminating at a different near-optimal solution.
I am using the same DDP convergence value that is given in multi-stage example. I didn't change it.
Question
I tried to run the multistage example provided in the GenX example system, but each simulation gives different results. I have attached pictures of the results from the power.csv file. Please check them.
As we can see in these pictures, when I run the example for the first attempt, the results of stage 3 (results_p3 power.csv) show that the battery is not contributing in Zone1 to resource allocation in MW per Zone. However, when I run the example again for the second time, the battery is contributing in Zone1 in the results of stage 3. Why are these results different?
Additional context
No response