Solving static output feedback problems by direct search optimization

Direct search methods are local optimization algorithms maximizing a possibly nonsmooth and nonconvex function using its values only (no first-order gradient or second-order Hessian information). More than one decade ago, N. J. Higham implemented some simple Matlab routines for direct search optimization, investigating questions on stability and accuracy of numerical algorithms in matrix computations. The purpose of this note is to report numerical experiments showing that with these routines a non-expert user can solve at moderate cost most of the static output feedback design problems available in the problem library COMPleib.