fun is doing odd things

I have defined a function fun

fun=@(refprice,r)sum((prices-refprice*ot+(exp(-r*t).*ot')').^2); 

and it works for any parameters refprice and r.

For example

fun(4.5,0.1)

ans =

1.1802e+04

>> when I put refprice0=4.5 and r0=0.1 i get

refprice0=4.5

refprice0 =

4.5000

>> r0=0.1

r0 =

0.1000

>> fun(refprice0,r0)

ans =

1.1802e+04

So it works.

But when I write

pars0=[refprice0,r0]

pars0 =

4.5000 0.1000

>> fun(pars0)

I get

Error using *

Incorrect dimensions for matrix multiplication. Check that the number of columns in the first matrix matches the number of

rows in the second matrix. To operate on each element of the matrix individually, use TIMES (.*) for elementwise

multiplication.

Error in analysedata>@(refprice,r)sum((prices-refprice*ot+(exp(-r*t).*ot')').^2) (line 13)

fun=@(refprice,r)sum((prices-refprice*ot+(exp(-r*t).*ot')').^2);

How can fun(refprice0,r0) work but not pars0=[refprice0,r0] and fun(pars0)???

[Torsten]

Question is answered here:

https://uk.mathworks.com/matlabcentral/answers/1987229-fmincon-is-not-working?s_tid=srchtitle

Answer 1

[VBBV]

fun is defined as anonymous function with two input parameters. In the 1st approach, you are defining both parameters as arguments to the anonymous function.

% 1st approach
prices = 10; ot = 20; t = 10;  % e,g values 
fun=@(refprice,r) sum((prices-refprice*ot+(exp(-r*t).*ot')').^2)

fun = function_handle with value:
    @(refprice,r)sum((prices-refprice*ot+(exp(-r*t).*ot')').^2)

fun(4.5,0.1)

ans = 5.2769e+03

% 2nd approach
refprice = 4.5;
r0 = 0.1;
pars0=[refprice,r0]

pars0 = 1×2

4.5000 0.1000

fun(pars0(1),pars0(2))

ans = 5.2769e+03

In the 2nd approach however, both parameters are defined as vector but still needs to be passed as 2 arguments as shown above. In both cases, it should yield same values.

In your case, you are passing the entire vector as single input parameter ignoring the other parameter. Since, the anonymous function contains multiplication operator, it expects a element wise multiplication when there is vector and throws an error when there are incorrect dimensions for the variables, which the error clearly states.

[John Hey]

Dear VBBV,

Many thanks for your answer.

John