Originally posted by: JDO
Hi everyone,
I'm trying to solve a program that a mathematical program that is linear and convex, except for a set of variables that I need the logarithm values of. I've tried to circumvent this issue by linearizing the logarithm functions via SOS2 constraints (as seen here).
However, doing so leads to an MIP that seems impossible to solve to integer feasibility (I have yet to find solutions regardless of the time limit). Anyone have suggestions of better ways to deal with concave functions such as the logarithm function?
Thanks,
JDO
#DecisionOptimization#OPLusingCPLEXOptimizer