Conic quadratic optimisation

[Fedmahn Kassad]

Hi all,

Is there any way to perform a Second Order Cone Programming (SOCP) optimisation with ISML, I've been through the list of optimisation types in the API and couldn't find any quadratic constraint.
The problem I need to solve is as follow:

Max	mTx
Subject To	‖GTx‖2≤a eTx=1 x≥0
Where	m: vector n x: vector n G: matrix n x n e: vector of unit 1 a: real number

I used to program conic optimisation with Mosek (https://docs.mosek.com/modeling-cookbook/cqo.html) and now I have to use ISML and couldn't find anything equivalent.

Thank you,
Fed

[jlo]

The problem can be solved using IMSL C# optimization class MinConNLP. It is necessary to reformulate the problem equivalently so it can be solved by MinConNLP. You may also need to use the MinConNLP.IGradient Interface to reduce numerical inaccuracies.