How to pass additional inputs to function in 'fminbnd'?

I have a function to optimize with respect to 'lambda':
function v = minfunction(lambda)
data = xlsread('data.xlsx');
data = data(:,2:3);
idx = xlsread('idx.xlsx');
v = 0;
for i = 1:size(idx,1)
v = v + max(0,(lambda*data(idx(i,1),1)+(1-lambda)*data(idx(i,1),2)-(lambda*data(idx(i,2),1)+(1-lambda)*data(idx(i,2),2))));
end
end
The optimization is of following form
lambda = fminbnd(@minfunction,0,1)
However, the function 'minfunction' asks for other arguments 'data' and 'idx'. These matrices are updated after each iteration. Currently, I am using excel data to pass these arguments. But I want to pass arguments 'data' and 'idx' in the 'minfunction' it self. If I am writing
function v = minfunction(data,idx,lambda)
and calling the optimizer as
lambda = fminbnd(@minfunction,0,1)
the MATLAB is showing error as too many input arguments.
Can someone help me with this? I am attaching the excel files with question containing idx and data.

 Accepted Answer

Do NOT call xlsread inside the function! Doing so will result in slow and inefficient code due to repeated file access.
The proper solution to your question is to define your function with as many arguments as you require:
function v = minfunction(data,idx,lambda)
and then pass it to the optimizer using an anonymous function:
idx = ...
lam = ...
v = fminbnd(@(t)minfunction(t,idx,lam), 0, 1);
Using this method the values of idx and lam are defined just once, and are then used for all iterations of the optimizer. You can read more this method by reading the MATLAB documentation:

4 Comments

Thanks a lot. I will check this.
Dear Stehphen,
It worked like a charm. I read the documentation and it resolved my issue.
Thanks again
Thanks Stephen, I needed the same answer.
lambda = fminbnd(@(lambda)minfunction(data,idx,lambda),0,1)

Sign in to comment.

More Answers (0)

Asked:

on 5 Feb 2018

Edited:

on 19 Mar 2020

Community Treasure Hunt

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

Start Hunting!