Excel Simplex LP vs Gurobi Optimizer
回答済みI'm trying to compare the results from using Gurobi vs. the Simplex LP method in Excel. I'm getting different results from the two methods. Even though they both solve to optimality, they both seem different. I understand that Simplex LP tries to find a global optimum, but Gurobi also does the same. Can anyone explain why there is a difference?
-
正式なコメント
-
Gurobi Staff
This post is more than three years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum. Or why not try Gurobot? -
-
-
Could you tell what exactly the difference is?
Are you 100% sure that you solve the same model in both cases? With Gurobi, you can use the write method to write the model to an LP file and analyze the file. If you are using the command line, you can set the ResultFile parameter. I don't know whether it is possible to somehow extract the LP model from Excel.
-1 -
-
Yes I can tell that the difference is off at about 4-5 around my decision variables. I'm 100% sure that I am solving the same model in both cases. I have noted the differences below. I'm not sure which solution solves a global optimum as I understood that Gurobi and Simplex both do so.
0 -
-
-
Note that Gurobi uses the Simplex algorithm as well.
It is possible that there are multiple optimal solution points for your optimization model. Did you compare the objective values provided by Excel and Gurobi? What is the difference between those?
Please also note that the tolerances used by Excel and Gurobi are probably slightly different which can lead to a slightly different solution point and ultimately a slightly different objective value. You could try adjusting Excel solver's tolerances to equal those of Gurobi for FeasibilityTol and OptimalityTol (or adjust Gurobi's tolerances to fit those of Excel).
0 -
-
Thank you for your help. I adjusted some of my decision variables and both methods managed to converge on the same values. I think it was the case where there were multiple optimal solution points in my model.
0 -
投稿コメントは受け付けていません。
コメント
5件のコメント