Re: Why the form of constraint affects the result of NMinimize?

*To*: mathgroup at smc.vnet.net*Subject*: [mg127615] Re: Why the form of constraint affects the result of NMinimize?*From*: Bill Rowe <readnews at sbcglobal.net>*Date*: Tue, 7 Aug 2012 03:03:47 -0400 (EDT)*Delivered-to*: l-mathgroup@mail-archive0.wolfram.com*Delivered-to*: l-mathgroup@wolfram.com*Delivered-to*: mathgroup-newout@smc.vnet.net*Delivered-to*: mathgroup-newsend@smc.vnet.net

On 8/6/12 at 4:39 AM, paperkite at gmail.com (paperkite rainyday) wrote: >I want to find the global minima of f[x1,x2,x3,x4] subject to >constraints, and I found that the results of >NMinimize[{f[x1,x2,x3,x4],x1>0 && x2>x3>0 && >x4>x2^2+x2*x3},{x1,x2,x3,x4}] and NMinimize[{f[x1,x2,x3,x4],x1>0 && >x2>x3 && x2>0 && x3>0 && x4>x2^2+x2*x3},{x1,x2,x3,x4}] are different >in mathematica 8. >Is there any difference between x2>x3>0 and x2>x3 && x2>0 && x3>0? >Why the form of constraint affects the result of NMinimize? You did not post details of your function f. Without those details, all I can do is guess. Most likely, the issue is related to the differences between machine precision and exact or symbolic math. To be any more accurate, precise or even know if this is on the right track requires knowing the details of the function you are trying to minimize.