Re: projected gradient
- To: mathgroup at christensen.cybernetics.net
- Subject: [mg1625] Re: projected gradient
- From: Scott.A.Hill at williams.edu (Lancelot)
- Date: Fri, 7 Jul 1995 00:13:03 -0400
- Organization: Williams College, Williamstown MA
In article <3tahig$6bb at news0.cybernetics.net>, James Albert Larson <larso171 at maroon.tc.umn.edu> wrote: >On 2 Jul 1995 22:59:29 GMT, >Filippo <teltesi25 at .polito.it> wrote: > >>I am an Electrical Engineering Student needing to find >>the constrained minimum of a function of several >>variables using the projected gradient algorithm. >>I would like to know if this algorithm is already >>been inplemented for Mathematica and should this be >>the case how to find it. > >Ans: no. The LinearProgramming and ConstrainedMin works with >constraints, but is only for linear objective function and linear >constraints. The only other optimization function in Mathematica is >FindMinimum -- the objective function can be nonlinear, but no constraints >are allowed. (There are also ConstrainedMax and FindMaximum -- the same as >ConstrainedMin and FindMinimum except maximizes rather than minimizes). I don't think there is a command called "FindMaximum", unless it's in a package somewhere that I don't have access to. No problem, really; FindMaximum[f] = FindMinimum[-f]. I wrote my own version of FindMinimum (called FindMin), since the former wasn't working for me and I wanted to see why it wouldn't work for me. I don't understand what you mean by "constrained" exactly (FindMinimum supposedly allows you to put bounds on the parameters searched, but I suppose that's not what is meant?), but if anyone would like the code and can tweak it to make it work this way, let me know. >Jim Larson