# Message Boards

Posted 9 years ago
3745 Views
|
0 Replies
|
0 Total Likes
|
 I tried using Method -> "QuadraticProgramming" in FindMinimum and got some strange error messages, which I illustrate with a simple problem. The strangest was that a constraint is required. In[21]:= FindMinimum[x^2 + y^2, {x, y}, Method -> "QuadraticProgramming"] During evaluation of In[21]:= FindMinimum::qpuncon: Method -> QuadraticProgramming cannot be used on an unconstrained optimization problem. >> Out[21]= FindMinimum[x^2 + y^2, {x, y}, Method -> "QuadraticProgramming"] In[24]:= FindMinimum[{x^2 + y^2, 1 <= 2}, {x, y}, Method -> "QuadraticProgramming"] During evaluation of In[24]:= FindMinimum::lpmp2: Warning: Method -> CLP is specified for a non-machine-precision problem;the problem will be converted into machine precision. Used Method -> Simplex for non-machine-number problem. >> Out[24]= {0., {x -> 0., y -> 0.}} In[26]:= FindMinimum[{x^2 + y^2, 1 <= 2}, {x, y}, Method -> "Simplex"] During evaluation of In[26]:= FindMinimum::bdmtd: Value of option Method -> Simplex is not Automatic, "Gradient", "ConjugateGradient", "InteriorPoint", "QuasiNewton", "Newton", "LinearProgramming", "QuadraticProgramming", or "LevenbergMarquardt". >> Out[26]= FindMinimum[{x^2 + y^2, 1 <= 2}, {x, y}, Method -> "Simplex"] In[25]:= FindMinimum[{x^2 + y^2, 1 <= 2}, {x, y}, Method -> "CLP"] During evaluation of In[25]:= FindMinimum::bdmtd: Value of option Method -> CLP is not Automatic, "Gradient", "ConjugateGradient", "InteriorPoint", "QuasiNewton", "Newton", "LinearProgramming", "QuadraticProgramming", or "LevenbergMarquardt". >> Out[25]= FindMinimum[{x^2 + y^2, 1 <= 2}, {x, y}, Method -> "CLP"]