http://www.mapleprimes.com/questions/95487-Constraints-Using-NLPSolve en-us 2026 Maplesoft, A Division of Waterloo Maple Inc. Maplesoft Document System Thu, 11 Jun 2026 18:24:02 GMT Thu, 11 Jun 2026 18:24:02 GMT The latest answers and comments added to the Question, Constraints using NLPSolve http://www.mapleprimes.com/images/mapleprimeswhite.jpg http://www.mapleprimes.com/questions/95487-Constraints-Using-NLPSolve http://www.mapleprimes.com/questions/95487-Constraints-Using-NLPSolve?ref=Feed:MaplePrimes:Constraints using NLPSolve:Comments#answer95495 <p>One cheap and dirty trick is to rewrite your procedure EIG as a piecewise, or with conditionals inside it. If any of the arguments are not "within bounds" then EIG can be edited to return instead some arbitrary <em>huge</em> value that you just make up. That way, the minimum shouldn't be found as any point outside bounds. (For maximization, you use a large negative value.)</p> <p>This can sometimes work OK when using nonlinearsimplex as the method, provided that you supply an initial point which is inside bounds (which you are already doing). The objective is no longer differentiable at the boundary, but in practice that aspect seems to not present many problems.</p> <p>The alternative of using sqp as the method, say, has problems for operator objectives like your EIG. See <a href="http://www.mapleprimes.com/posts/95113-Optimization-Workaround-For-Operator-Form">here</a>, which shows a workaround that involves handling of the objectivegradient but not of the constraintjacobian.</p> <p>One cheap and dirty trick is to rewrite your procedure EIG as a piecewise, or with conditionals inside it. If any of the arguments are not "within bounds" then EIG can be edited to return instead some arbitrary <em>huge</em> value that you just make up. That way, the minimum shouldn't be found as any point outside bounds. (For maximization, you use a large negative value.)</p> <p>This can sometimes work OK when using nonlinearsimplex as the method, provided that you supply an initial point which is inside bounds (which you are already doing). The objective is no longer differentiable at the boundary, but in practice that aspect seems to not present many problems.</p> <p>The alternative of using sqp as the method, say, has problems for operator objectives like your EIG. See <a href="http://www.mapleprimes.com/posts/95113-Optimization-Workaround-For-Operator-Form">here</a>, which shows a workaround that involves handling of the objectivegradient but not of the constraintjacobian.</p> 95495 Wed, 28 Jul 2010 00:17:26 Z pagan pagan http://www.mapleprimes.com/questions/95487-Constraints-Using-NLPSolve?ref=Feed:MaplePrimes:Constraints using NLPSolve:Comments#answer95504 <p>Have you tried Search command from <a href="http://www.maplesoft.com/applications/view.aspx?SID=87637">DirectSearch</a> package ?</p> <p>Have you tried Search command from <a href="http://www.maplesoft.com/applications/view.aspx?SID=87637">DirectSearch</a> package ?</p> 95504 Wed, 28 Jul 2010 16:01:32 Z mois mois