I am looking for a good method to reduce the number of scenarios in problems with a large number of uncertainty parameters.
I implemented some of these methods in MATLAB, but unfortunately the number of scenarios does not decrease to the value I want. For example, the 8,000 scenarios generated are reduced to only 1,500 scenarios, which is still a large number.
1. Do you know the suitable method for scenario reduction? Please introduce.
2. Do you have a good resource for learning "ScenRed" method in GAMS software? Please introduce.
Dear Farzan Ghasemi,
I suggest you to see links and attached files on topic.
https://www.sciencedirect.com/science/article/pii/S0098135415001623
https://www.aiche.org/conferences/aiche-annual-meeting/2013/proceeding/paper/696a-oscar-optimal-scenario-reduction-tool-based-on-distance-uncertainty-distribution-and-output-1
https://lib.dr.iastate.edu/cgi/viewcontent.cgi?article=2802&context=etd
https://arxiv.org/pdf/1908.11486.pdf
Article Optimal Scenario Reduction Framework based on Distance of Un...
Best regards
Dear Farzin
I understand that what you call scenarios are alternatives but I don't see their relationship with uncertain data.
Please explain, or put an example
Reduction of alternatives can be done using either Principal Components Analysis, a statistical technique, or Linear Programming or SIMUS.
In these last two, you simply specify the number of scenarios you want in a final set of alternatives. The size of this set depends on the number of criteria. If there are a large number of criteria, it could be that the size of the alternatives set is not enough to satisfy all criteria simultaneously and vice versa
Dear Farzin Ghasemi-Olanlari,
You can use the below link to access a tutorial that uses meta-heuristic methods to scenario reductions by adopting algorithms like K_Means.
https://faradars.org/courses/fvrp106-automatic-clustering-using-evolutionary-algorithms-and-metaheuristics
Best Regards.
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning