main routine to call -- eigopt_constrained.m Type >> help eigopt_constrained for information on input and output arguments fpsa.m - to compute pseudospectral abscissa fpsr.m - to compute pseudospectral radius robustcontrol.m - to solve the robust stabilization problem structuredesign_bilinear.m - to solve the structural design problem (bilinear matrix-valued function) structuredesign.m - to solve the structural design problem (affine matrix-valued function) %%%%%%%%%%%%% A sample run to compute the epsilon-pseudospectral abscissa of a matrix A: %%%%%%%%%%%%%%% >> pars.A = randn(100); >> pars.AA = A'*A; >> pars.epsln = 10^-1; >> pars.gamma = 2; >> eigvals = eig(pars.A); >> [m,ind] = max(real(eigvals)); >> pars.x0 = [real(eigvals(ind)); imag(eigvals(ind))]; >> [f,x,xvec,fvec] = eigopt_constrained('fpsa',[1 0]',pars); %%%%%%%%%%%%%%% A sample run for robust stabilization problem. See [1] Section 8, in particular problem (8.4) %%%%%%%%%%%%%%% >> load robustcontrol >> c = [-norm(pars.K{2}) -norm(pars.K{3})]'; >> [fobj,x,xvec,fvec] = eigopt_constrained('robustcontrol',c,pars); %%%%%%%%%%%%%% A sample run for structural design problem See [1] Section 8, problem (8.5) %%%%%%%%%%%%%% >> load structuredesign_bilinear >> c = [-1 -1]'; >> [fobj,x,xvec,fvec] = eigopt_constrained('structuredesign_bilinear',c,pars); %%%%%%%%%%%%%% REFERENCES: [1] - E. Mengi. A Support function based algorithm for optimization with eigenvalue constraints. Submitted to SIOPT.