NOTICE: developerWorks Community will be offline May 29-30, 2015 while we upgrade to the latest version of IBM Connections. For more information, read our upgrade FAQ.
This topic has been locked.
2 replies Latest Post - 2013-01-15T21:53:47Z by J2C0_Mark_Bloomfield
Pinned topic feasOpt in the MATLAB API
Answered question This question has been answered.
Unanswered question This question has not been answered yet.
In the CPLEX MATLAB API it only seems possible to pass linear constraints to the feasOpt call. Is there any way of to pass quadratic constraints too? I think this capability is available in other APIs, but am keen to know whether this can be done in MATLAB. Thanks!
Updated on 2013-01-15T21:53:47Z at 2013-01-15T21:53:47Z by J2C0_Mark_Bloomfield
SystemAdmin 110000D4XK7929 PostsACCEPTED ANSWER
Re: feasOpt in the MATLAB API2013-01-15T12:01:08Z in response to J2C0_Mark_BloomfieldYou are right. In the matlab API feasopt() can only be invoked for linear problems. Quadratic constraints are not supported here. I have filed a user wish that asks for this feature.
To get feasopt for quadratic constraints you would have to manually create the model that feasopt would otherwise create: Add additional variables to the quadratic constraints that make the constraints always feasible and penalize the variables if they are non-zero. Then minimize the penalty.