2009 Poster Sessions : Real-time Convex Optimization with Applications

Student Name : Jacob Mattingley, Yang Wang, Argyrios Zymnis
Advisor : Stephen Boyd
Research Areas: Information Systems
Abstract:
We will present various applications of fast real-time convex optimization. We will show examples of how convex optimization can be applied to control of mechanical systems, network utility maximization, processor scheduling, estimation, and fault detection.
Recent results show that these problems can be solved extremely efficiently, with computation times often in milliseconds or microseconds for small and medium sized problems. This allows convex optimization to be embedded in real-time fully automated systems. We will also describe an automatic code generation system for real-time embedded convex optimization, and we will show timing results from a preliminary implementation, built on a Python-based modeling framework CVXMOD.

Bio:
Jacob Mattingley, Yang Wang, and Argyrios Zymnis are all PhD students of Stephen Boyd in the Information Systems Laboratory at Stanford University. Their research interests include convex optimization with applications to automatic control, signal processing, and machine learning.