The problem of not supporting exponential functions when using Matlab + Yalmip+ Gurobi to establish an optimization problem

[家国曾]

Hi,

When I use MATLAB + Gurobi + YALMIP to build an optimization model and include exponential functions or log functions in the constraints, I get the following error when solving:

solvertime: 0
info: 'Solver not applicable (gurobi does not support exponential cones)'
problem: -4
yalmiptime: 0.1500

There are no further error messages. After contacting Gurobi’s official support, they said the issue lies with YALMIP, which is unable to convert the model into a form supported by Gurobi.

How can I resolve this issue? Could it be due to an incompatible version of YALMIP?

Thanks for your help!

[Johan Löfberg]

It's available in the develop branch.

Note though, using exponentials etc in gurobi means you are solving very demanding global optimization problems through approximations and spatial branch and bound. If the model is convex (for relaxed integers) you are likely much better off using mosek and the exponential cone framework