<meta http-equiv="Content-Type" content="text/html; charset=utf-8" /><P>Hi,</P>
<P>Firstly,thank you for your attention to this E-mail.On the website </P>
<P><A href="http://www.coin-or.org/Ipopt/documentation/node105.html">http://www.coin-or.org/Ipopt/documentation/node105.html</A> presents how to reset the keyword </P>
<P>"mehrotra_algorithm". If set to yes, Ipopt runs as Mehrotra's predictor-corrector algorithm. </P>
<P>This automatically disables the line search, and chooses the adaptive mu strategy with the </P>
<P>"probing" oracle, and uses "corrector_type=affine" without any safeguards,you should not set </P>
<P>any of those options explicitly in addition.In my opinion,resetting option "mehrotra_algorithm yes"</P>
<P>equals to resetting options "mu_strategy adaptive" + "mu_oracle probing" + "corrector_type affine",</P>
<P>however the different iteration processing appeared when solving nonlinear optimization problems under</P>
<P>aforementioned two configurations,the former configuration automatically disables the line search and </P>
<P>manifests worse performance,and the latter configuration still enables the line search,such as resetting</P>
<P>option max_soc nonzero number.Can someone explain this phenomenon for me? Which is the correct configuraton</P>
<P>that brings Mehrotra's predictor-corrector algorithm into effect? </P>
<P>There are three options of keyword "corrector_type":none,affine,primal-dual.What's the difference between</P>
<P>affine corrector and primal-dual corrector? In the paper "Adaptive Barrier Strategies for Nonlinear Interior </P>
<P>Methods" can be found on website <A href="http://epubs.siam.org/doi/pdf/10.1137/060649513">http://epubs.siam.org/doi/pdf/10.1137/060649513</A> ,part 7 "Corrector Steps" </P>
<P>only presents one method to calculate corrector steps.</P>
<P>Sincerely.</P>
<P>From Wei</P>
<P><BR> </P>