Solving MILP faster (Gurobi-CVX)
I’m trying to solve a MILP where Gurobi (within CVX) takes more than a week to solve. Can anyone suggest me how it is possible to instruct Gurobi through CVX settings to terminate when the cost is below a certain threshold? I intend to compute the threshold from upper bound and lower bound of cost.
Is there any other way to solve the optimization faster? Does anyone recommend use of MIPGap through CVX settings? If so it would be great to have a direction?
-
Solvers — CVX Users' Guide (cvxr.com) has instructions on how to pass parameter settings to Gurobi.
To create a new custom setting for the current solver, use this syntax:
cvx_solver_settings( '{name}', {value} )
0 -
Hi, Philip Taffet thanks for your suggestion. I tried to use this. But not sure what "value" to use. Can you suggest me how to use MIPGap through CVX also? or is there any other way to solve such problems faster (just in case you have had similar experience) ?
0
Please sign in to leave a comment.
Comments
2 comments