<div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><div id="isForwardContent"><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial">Thank you for your reply! Setting "bound_relax_factor" to 0 did help avoid the error.</div><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><br></div><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial">Moreover, I'd like to know if disabling&nbsp;bounds relaxation ("bound_relax_factor" = 0) would reduce the algorithm efficiency?&nbsp;</div><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial">After setting "bound_relax_factor" to 0, IPOPT could hardly solve the model.</div><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><span class="Apple-style-span" style="white-space: pre-wrap; "><br></span></div><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><span class="Apple-style-span" style="white-space: pre-wrap; ">Thank&nbsp;you&nbsp;very&nbsp;much!&nbsp;Looking&nbsp;forward&nbsp;to&nbsp;your&nbsp;reply.</span></div><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><span class="Apple-style-span" style="white-space: pre-wrap; ">&nbsp;</span></div><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><span class="Apple-style-span" style="white-space: pre-wrap; ">Cavern</span></div><div style="line-height:1.7;color:#000000;font-size:14px;font-family:arial"><pre><span class="Apple-style-span" style="white-space: normal;"><br></span>At&nbsp;2012-06-13&nbsp;23:52:00,"Stefan&nbsp;Vigerske"&nbsp;&lt;<a href="mailto:stefan@math.hu-berlin.de">stefan@math.hu-berlin.de</a>&gt;&nbsp;wrote:
&gt;Hi,
&gt;
&gt;Ipopt&nbsp;relaxes&nbsp;the&nbsp;bounds&nbsp;slightly,&nbsp;probably&nbsp;too&nbsp;much&nbsp;for&nbsp;your&nbsp;case.
&gt;You&nbsp;can&nbsp;disable&nbsp;this&nbsp;by&nbsp;setting&nbsp;the&nbsp;option&nbsp;bounds_relax_factor&nbsp;to&nbsp;0.
&gt;
&gt;However,&nbsp;a&nbsp;bound&nbsp;of&nbsp;0.00000001&nbsp;is&nbsp;not&nbsp;much&nbsp;better&nbsp;than&nbsp;0.&nbsp;Even&nbsp;though&nbsp;
&gt;the&nbsp;log&nbsp;can&nbsp;be&nbsp;evaluated&nbsp;there,&nbsp;the&nbsp;numerical&nbsp;behavior&nbsp;of&nbsp;the&nbsp;model&nbsp;
&gt;around&nbsp;this&nbsp;value&nbsp;is&nbsp;bad.
&gt;
&gt;Stefan
&gt;
&gt;On&nbsp;13.06.2012&nbsp;16:03,&nbsp;Õų¿&nbsp;wrote:
&gt;&gt;&nbsp;Hi&nbsp;all:
&gt;&gt;&nbsp;When&nbsp;I&nbsp;set&nbsp;"var&nbsp;a&gt;=0.00000001&nbsp;default&nbsp;0.001"&nbsp;and&nbsp;solve&nbsp;the&nbsp;model&nbsp;by&nbsp;IPOPT&nbsp;with&nbsp;AMPL,&nbsp;IPOPT&nbsp;would&nbsp;return&nbsp;the&nbsp;error&nbsp;information&nbsp;"can't&nbsp;evaluate&nbsp;log(-3.05e-10)"&nbsp;(base&nbsp;number&nbsp;is&nbsp;very&nbsp;slightly&nbsp;less&nbsp;than&nbsp;zero)&nbsp;in&nbsp;equations&nbsp;including&nbsp;the&nbsp;"log(a)"&nbsp;term&nbsp;sometimes.&nbsp;Similarly,&nbsp;it&nbsp;would&nbsp;return&nbsp;"can't&nbsp;evaluate&nbsp;power(1.00035,&nbsp;10e08)"&nbsp;(base&nbsp;number&nbsp;is&nbsp;very&nbsp;slightly&nbsp;more&nbsp;than&nbsp;1)&nbsp;in&nbsp;equations&nbsp;including&nbsp;the&nbsp;"[1/(1+a)]^(10e08)"&nbsp;term.
&gt;&gt;
&gt;&gt;
&gt;&gt;&nbsp;So&nbsp;why&nbsp;would&nbsp;this&nbsp;phenomenon&nbsp;happen?&nbsp;It&nbsp;seems&nbsp;that&nbsp;the&nbsp;lower&nbsp;bound&nbsp;of&nbsp;"a"&nbsp;didn't&nbsp;take&nbsp;effect.&nbsp;How&nbsp;can&nbsp;I&nbsp;avoid&nbsp;this&nbsp;error?
&gt;&gt;
&gt;&gt;
&gt;&gt;&nbsp;Besides,&nbsp;if&nbsp;there&nbsp;are&nbsp;some&nbsp;terms&nbsp;like&nbsp;exp(a)&nbsp;or&nbsp;a^b&nbsp;(a,b&nbsp;are&nbsp;variables)&nbsp;in&nbsp;AMPL&nbsp;equations,&nbsp;do&nbsp;we&nbsp;always&nbsp;need&nbsp;to&nbsp;change&nbsp;them&nbsp;to&nbsp;the&nbsp;logarithm&nbsp;form&nbsp;(e.g.&nbsp;a^b&nbsp;becomes&nbsp;b*log(a)&nbsp;)?&nbsp;Does&nbsp;this&nbsp;reformulation&nbsp;do&nbsp;make&nbsp;the&nbsp;solving&nbsp;process&nbsp;more&nbsp;efficient?
&gt;&gt;
&gt;&gt;
&gt;&gt;&nbsp;Thank&nbsp;you&nbsp;very&nbsp;much!&nbsp;Looking&nbsp;forward&nbsp;to&nbsp;your&nbsp;reply.
&gt;&gt;
&gt;&gt;
&gt;&gt;&nbsp;Yours
&gt;&gt;&nbsp;Cavern
&gt;&gt;
&gt;&gt;
&gt;&gt;
&gt;&gt;&nbsp;_______________________________________________
&gt;&gt;&nbsp;Ipopt&nbsp;mailing&nbsp;list
&gt;&gt;&nbsp;<a href="mailto:Ipopt@list.coin-or.org">Ipopt@list.coin-or.org</a>
&gt;&gt;&nbsp;http://list.coin-or.org/mailman/listinfo/ipopt
</pre></div></div><br><br><span title="neteasefooter"><span id="netease_mail_footer"></span></span></div></div><br><br><span title="neteasefooter"><span id="netease_mail_footer"></span></span>