How can I apply a filter using fminunc?

1 次查看(过去 30 天)
I have a function f(x) to minimize using fminunc, but I want to apply a filter to x after every iteration. This is a simple example, just to explain:
x=4;
func=@(x)myfunct(x)
options=optimoptions('fminunc','SpecifyObjectiveGradient',true,'Display','iter');
x0=2;
[best_x, fval, exitflag, output] = fminunc(func,x0,options)
% myfunction
function [OF,grad_OF]=myfunct(x)
OF=3*x^2+2*x;
grad_OF=6*x+2;
In this case the solver starts from x0, until it finds the minimum; I would that after every iteration, my x0 is x0=2*best_x (or x0=best_x - 2....) Now it is not important the specific relation, but how can I use and modify the best_x in every iteration, using it as the new starting point in fminunc solver.

采纳的回答

Matt J
Matt J 2017-11-25
编辑:Matt J 2017-11-25
You can run fminunc in a loop with a single iteration in each pass (by setting MaxIter=1). Within the loop, you can also apply your filter.
options=optimoptions('fminunc','SpecifyObjectiveGradient',true,...
'Display','iter','MaxIter',1);
best_x=x0;
for i=1:N
[best_x, fval, exitflag, output] = fminunc(func,best_x,options)
best_x=myFilter(best_x);
end

更多回答(0 个)

类别

Help CenterFile Exchange 中查找有关 Solver Outputs and Iterative Display 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by