Multi objective optimization - non dominant solution
AnsweredHi Jaromil,
I had a question regarding a bi objective problem that I am solving. I have two objectives: objective 1 minimizes total travel time and objective 2 minimizes total unmet demand. I am using the reltol on objective 1 and allowing to objective 2 to degrade the value by giving a specified relative amount. I am setting the reltol values as 1%,2%,3%,4% and 5% to set the amount for relative degradation. Since the dataset for my problem is very large, I do not want to exhaustively find all optimal solution. I am only interested in finding the Pareto Frontier, for which I am finding using Reltol.
My question was - Are all these value given by Gurobi, the non-dominant solutions for this bi-objective function? When I plot the values in a scatter plot, it does form an inverse relationship graph( as shown in the below image) so it does look like a pareto frontier for this optimization.

I also wanted to added that I am solving the bi objective problem by hierarchical lexicographic method. The priority of objective 1 is 2 and the priority of objective 2 is 1. Here is the piece of code for your reference :
m.setObjectiveN(Obj1, index = 0, priority = 2, name = "Obj1", reltol = 0.01)
m.setObjectiveN(Obj2, index = 1, priority = 1, name = "Obj2")
Thanks,
Purnima
-
Official comment
This post is more than two 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 try Gurobot, our chatbot interface offering instant, expert-level support. -
This is a cross-post and is already addressed.
0
Post is closed for comments.
Comments
2 comments