These 3 codes when run give error

I have 3 codes as given in attachment. When I run the "main.m", it gives the following error:

Index in position 1 exceeds array bounds. Index must not exceed 1.

Error in ArGeo (line 17)

cosd(b(2, :)) .* cosd(b(1, :)); % x-component

Error in main>@(b)ArGeo(b,u,Noise)+penaltyTerm(b,u) (line 25)

[time,gBest,gBestScore]=WHO(20,100,lb,ub,dim,@(b) ArGeo(b,u,Noise)+penaltyTerm(b, u));

Error in WHO (line 47)

group(i).cost=fobj(group(i).pos);

Error in main (line 25)

[time,gBest,gBestScore]=WHO(20,100,lb,ub,dim,@(b) ArGeo(b,u,Noise)+penaltyTerm(b, u));

>>

Answer 1

[Voss]

In WHO.m lines 45-48, you have

for i=1:Nfoal

group(i).pos=lb+rand(1,dim).*(ub-lb);

group(i).cost=fobj(group(i).pos);

end

where dim is 2, and lb and ub are each 1xdim vectors. Thus, group(i).pos is created as a 1xdim vector on line 46, and the error happens after fobj (which is @(b) ArGeo(b,u,Noise)+penaltyTerm(b, u)) is called on line 47, because inside ArGeo.m, input b, which is group(i).pos in WHO.m, is expected to have at least 2 rows:

kb = pi * [

cosd(b(2, :)) .* cosd(b(1, :)); % x-component

sind(b(2, :)) .* cosd(b(1, :)); % y-component

sind(b(1, :)) % z-component

];

That's the reason for the error.

As far as how to fix it, I suspect that lines 26-29 of WHO.m are not doing what's intended:

if size(ub,1)==1

ub=ones(1,dim).*ub;

lb=ones(1,dim).*lb;

end

Since ub and lb are each 1xdim vectors, you are multiplying each element of ub and lb by 1, which leaves them unchanged. Instead, I guess you want to replicate them so that they have exactly 2 rows (not dim rows), which could be done like:

if size(ub,1)==1

ub=ones(2,1).*ub;

lb=ones(2,1).*lb;

end

Then they are of size 2xdim, which seems like what's intended.

[Sadiq Akbar]

@Voss

Thanks a lot for your kind response. I did that changes and ran it but now it gives the following error:

Index exceeds the number of array elements. Index must not exceed 4.

Error in exchange (line 16)

bestgroup=Stallion(i).group(index);

Error in WHO (line 77)

Stallion=exchange(Stallion);

Error in main (line 25)

[time,gBest,gBestScore]=WHO(20,100,lb,ub,dim,@(b) ArGeo(b,u,Noise)+penaltyTerm(b, u));

>>

[Voss]

You're welcome! Since the original error is solved, please "Accept" this answer. Thanks!

The new error means that i or index is greater than 4 on the line

bestgroup=Stallion(i).group(index);

and that's a problem because (if it's i) Stallion has only 4 elements or (if it's index) Stallion(i).group has only four elements.

You can set a breakpoint (https://www.mathworks.com/help/matlab/matlab_prog/set-breakpoints.html) at that line and check the value of the variables.

If you're not sure what to do to fix it, please upload the file exchange.m (and any other m-files it needs to run), and I'll take a look.

[Sadiq Akbar]

Thanks a lot for your kind response. Sorry I had forgotten to upload the supporting files for WHO. Here are they attached now.

Important: Actually most of time I keep one row in all u vetros. So once again I have replaced the u vectors by a single row instead of two rows because when u has one row, then may be you solve it quickly. So I uploaded the main.m file also.

I hope now you will solve it very quickly because its easy for you but very difficulat for people like me.

[Voss]

Regarding the error

Index exceeds the number of array elements. Index must not exceed 4.

Error in exchange (line 16)

bestgroup=Stallion(i).group(index);

Error in WHO (line 77)

Stallion=exchange(Stallion);

Error in main (line 25)

[time,gBest,gBestScore]=WHO(20,100,lb,ub,dim,@(b) ArGeo(b,u,Noise)+penaltyTerm(b, u));

the first time exchange is run, when i is 1, this line

[value,index]=min([Stallion(i).group.cost])

produces an index value of 7. The array

[Stallion(i).group.cost]

is of size 1x12 because Stallion(1).group is a 1x4 struct array and each element of it contains a cost field which is 1x3. So, it's not appropriate to try to index Stallion(i).group with the index of the minimum element in [Stallion(i).group.cost].

Furthermore, Stallion(1).cost is a 1x3 vector, so comparing it with value like this

if value<Stallion(i).cost

means that the if condition is true only when value is less than all elements of Stallion(i).cost. Is that what you want to do?

In general, I recommend putting breakpoints in your code and making sure the sizes of various variables make sense and are as intended, as the code runs.

https://www.mathworks.com/help/matlab/matlab_prog/debugging-process-and-features.html#mw_aac9fccc-eee5-48d5-9c1e-36a598251c93

[Sadiq Akbar]

Actually the I don't understand the coding of WHO algorithm because I have downloaded it from the Mathworks site. However, the ArGeo.m is my function. Normally I have been using the u vector as having a single row but since it was giving error, that's why I put a semicolon inside u and made two rows in u. I thought that it will remve the error but in vain.

Further, if you run the ArGeo.m as a script alone, it works. for example I attach it here, you run it, it works but I don't don't know why it gives error with main.m?

[Voss]

OK. Since WHO is a File Exchange entry, forget what I said in my answer, and go back to the original code you downloaded from the File Exchange.

Your ArGeo assumes u and b have (at least 2 rows), so the u defined in your main needs to have (at least) 2 rows:

u_sets = {[-33 -40; 33 40], [-50 -60 -70; 50 60 70], [-20 -30 -42 -52; 20 30 42 52]};

I think ub and lb need to be the same size as u, so in main:

[n,dim]=size(u);

lb = -90*ones(n,dim);

ub = 90*ones(n,dim);

And I think the cost assigned to each Stallion needs to be a scalar, which your ArGeo returns a scalar but then a penalty term is added, which is not necessarily a scalar, so in main:

penalty = 100 * sum((x - desiredResult).^2,'all');

% ^^^^^^ sum over all elements

With those three changes in main, the code runs, at least past the WHO part, until the displaying the results part, where an error with concatentating a character vector with num2str(u) occurs, because u has multiple rows. You can use mat2str(u) instead, to get around that. After that, the code completes.

As to whether the results make any sense, I haven't the foggiest idea, since I have no references for the WHO algorithm or any explanation of the assumptions made in that code.

I've attached an updated main.m with all of those changes.

[Sadiq Akbar]

@Voss

Thanks a lot for your kind response. Yes, it runs but gives absolutely wrong results. When I run the main, it should give me nearly the values of u. As you can see the 1st set of u inside main is having four values i.e.,

[-33 -40;33 40], so after running the main, it should return me these values i.e., [-33 -40;33 40]. Likewise, the 2nd set of u inside main is : [-50 -60 -70;50 60 70], so after running the main, it should retun me these values and so on. But when I ran your modified main, the code runs but it gives me these values:

[0.00032681 0.00032681] instead of [-33 -40;33 40]

[-0.0033306 -0.0033306] instead of [-50 -60 -70;50 60 70]

[0.0096308 0.0096308] instead of [-20 -30 -42 -52;20 30 42 52]

So I don't want to run just the code but rather I want to get these results.

[Voss]

I don't know.

[Sadiq Akbar]

@Voss

Dear Voss my problem is not solved but looking at your time that you spared for my problem, it's precious, so I accept your answer.

Regards,