Branch and bound computation of the minimum norm of a linear fractional transformation over a structured set

The minimum norm of a linear fractional transformation (LFT) over a structured set is computed using a branch and bound algorithm. This is a global optimization problem caused by the possibility of local minima. Several computationally efficient lower bounds for the minimum norm of the LFT are developed, and it is demonstrated that the success of the optimization, as measured by time-to-converge, largely depends on the quality of these bounds.

[1]  Andrew Packard,et al.  The complex structured singular value , 1993, Autom..

[2]  Stephen P. Boyd,et al.  Global Optimization in Control System Analysis and Design , 1992 .

[3]  Peter M. Young,et al.  Mixed μ problems and branch and bound techniques , 1997 .

[4]  Stephen P. Boyd,et al.  Branch and bound algorithm for computing the minimum stability degree of parameter-dependent linear systems , 1991, International Journal of Robust and Nonlinear Control.

[5]  Stephen P. Boyd,et al.  A branch and bound methodology for matrix polytope stability problems arising in power systems , 1990, 29th IEEE Conference on Decision and Control.

[6]  J. Doyle,et al.  Robust and optimal control , 1995, Proceedings of 35th IEEE Conference on Decision and Control.

[7]  J. D. Sipila,et al.  Optimal structure design using branch and bound , 1999, Proceedings of the 1999 American Control Conference (Cat. No. 99CH36251).