Skip to Main Content
This article shows the potential for convex optimization methods to be much more widely used in signal processing. In particular, automatic code generation makes it easier to create convex optimization solvers that are made much faster by being designed for a specific problem family. The disciplined convex programming framework that has been shown useful in transforming problems to a standard form may be extended to create solvers themselves. Much work remains to be done in exploring the capabilities and limitations of automatic code generation. As computing power increases, and as automatic code generation improves, the authors expect convex optimization solvers to be found more and more often in real-time signal processing applications.