In this case we use the sample problem given to us by Schilders.
>> !cp examples/simschur/*.m . >> [A,B] = noisy; >> parameters(A,B); >> Y0 = guess; >> [fn,Yn] = sg_min(Y0,'prcg','euclidean'); iter grad F(Y) flops 0 2.205361e-01 1.687421e-02 56493 1 1.087502e-01 8.329681e-03 1537489 2 5.834580e-02 5.947527e-03 3045842 3 3.954742e-02 5.322286e-03 4243901 4 1.920065e-02 5.016544e-03 5700349 5 9.449972e-03 4.938962e-03 7062826 6 4.304992e-03 4.920311e-03 8509245 7 2.649005e-03 4.916373e-03 9982763 8 1.839592e-03 4.914510e-03 11492687 9 1.162756e-03 4.913601e-03 12952876 10 5.450777e-04 4.913326e-03 14401149 11 2.461091e-04 4.913273e-03 15848402 12 1.463300e-04 4.913261e-03 17354270 13 7.851036e-05 4.913257e-03 18762063 14 5.400429e-05 4.913256e-03 20205661 15 2.786051e-05 4.913255e-03 21689844 16 1.570592e-05 4.913255e-03 23166944 17 8.390343e-06 4.913255e-03 24528843 18 4.060847e-06 4.913255e-03 25901423 19 2.047984e-06 4.913255e-03 27271928 20 1.161407e-06 4.913255e-03 28481228 21 7.367792e-07 4.913255e-03 29776935
22 4.640484e-07 4.913255e-03 31533888 23 2.206626e-07 4.913255e-03 33245809 24 1.467168e-07 4.913255e-03 35084169 25 7.946751e-08 4.913255e-03 36975859 26 4.631213e-08 4.913255e-03 38934071 27 2.119895e-08 4.913255e-03 40761932
Figure 9.5 shows the convergence curve for this run.
