Solving Large Sum-of-Squares Optimization Problems in Control by Exploiting the Parallel Structure of Polyas Algorithm

To date, the CSCL has retained a graduate student (Reza Kamyar). Reza and the PI have produced parallel code for solving linear robust analysis and control problems. Using clustercomputing and supercomputing resources at IIT and ANL, he has analyzed dynamic systems with uncertainties on the simplex and with up to 100 states. He has extended this code to systems with uncertainties on the hypercube. He has published 2 conference papers and 1 journal paper (under revision). The student Reza Kamyar will transition to ASU from IIT. At ASU, Reza and the PI will continue to advance algorithms and code for analysis and control of large-scale systems using advanced computational resources. Specifically, Reza will extend his code to consider analysis of nonlinear systems and optimal control of uncertain linear systems. He will also attempt to adapt his work to other advanced computing platforms such as GPU computing.