A CONIC APPROACH FOR SEPARABLE CONVEX OPTIMIZATION Two classes of structured convex optimization problems known as geometric optimization and l_p-norm optimization have been recently studied in the framework of conic optimization, relying on the definition of dedicated convex cones. In this talk, we present a generalization of these cones that allows us to model a wide class of separable convex problems. We also investigate the development of interior-point methods for this class of problems using the theory of self-concordant barriers.