How to call objective function in outputfcn or plotfcn during optimization?

1 view (last 30 days)
I want to minimize a complicated objective function. It returns 1. objective function value and 2. some fit data (like the fitted line from a ls regression). I want to use the 'OutputFcn' or 'PlotFcn' argument to draw the fitted line given the paramters at the current iteration. Unfortunately the fields in optimValues do not contain the obj_function, so I don't know how to do it. Can anyone help?
As a MWE consider the code below which does the least squares regression numerically.
%% Random Data
x = 1:.1:20;
x = x';
e = randn(numel(x),1)*10;
y = 1 + 3*x + e;
%% Optimization
eparam = fminunc(@(param) obj_fun(param, x, y), [0 0]);
%% Plot I want drawn at every iteration
[sum_of_squared_residuals,y_pred] = obj_fun(eparam,x,y);
plot(x,y)
hold on
plot(x, y_pred)
%% Objective function
function [sum_of_squared_residuals,y_pred] = obj_fun(param,x,y)
intrcpt = param(1);
slope = param(2);
y_pred = intrcpt + slope*x;
squared_residuals = (y-y_pred).^2;
sum_of_squared_residuals = sum(squared_residuals);
end

Accepted Answer

Matt J
Matt J on 3 May 2021
Unfortunately the fields in optimValues do not contain the obj_function
It does. It's the fval field.
  6 Comments

Sign in to comment.

More Answers (0)

Categories

Find more on Optimization in Help Center and File Exchange

Community Treasure Hunt

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

Start Hunting!