Problem using the Fuzzy Logic Toolbox

I'm trying to use the Fuzzy Logic Toolbox, and I'm having some troubles that I have no idea how to get around. I use the evalfis function which runs ok until it calls evalfismex.

From this point, it says "MATLAB function '' does not exist!"

Any thoughts?

Answer 1

[Walter Roberson]

Does it name the function that does not exist? Could you show us the traceback?

Answer 2

[Trevor]

No, it doesn't name the function. I have determined that it is a problem with evalfismex though, as I have put a breakpoint there and as soon as I try to step into it, that error comes up.

I'm not sure what you mean by the traceback though. Can you clarify?

[Walter Roberson]

I see now the '' in the middle of the error message, in the place the function name would go.

Try

which -all evalfismex

and see if it comes up with something. If it doesn't, then I would expect a complaint about evalfismex not existing. If it does come up with something, then either that evalfismex is corrupted or else it calls back to something that isn't defined for some reason.

It appears that the source code for evalfismex is included as toolbox\fuzzy\fuzzy\src\evalfismex.c

(reference: http://www.mathworks.com/matlabcentral/newsreader/view_thread/139525 )

[Trevor]

I tried it and it came up with something close to the path you gave:

C:\Program Files\MATLAB\R2010b\toolbox\fuzzy\fuzzy\evalfismex.mexw64

I'm not sure if the discrepancy is why it is coming up with an error.

[Walter Roberson]

There is no real discrepancy there: the name I gave earlier was for the C source. With the C source, in theory you should be able to rebuild the .mexw64 -- though you might want to take a look at it and see if it appears to have any calls back to the MATLAB engine (if so then if one of the routines it needs does not exist then rebuilding it would not help.)

I would suggest you toss this over to TMW support as they have access to the bug database.

[Trevor]

I tried recompiling the .mexw64 file, but it doesn't find a compiler that I can use. Unfortunately, I'm remotely accessing the program from my university, so I can't even really install one.

Thanks for all your help, I'll see what TMW support can do for me.

Answer 3

[Aleksandar Mihaylov]

I know this is an old post, but the problem has a simple solution: Make sure that ALL membership functions, along with the methods for and, or, defuzz, implication and aggregation are assigned using single quotes: i.e. 'prod' not prod. Further: Right: fis = addmf(fis, 'input', 1, '1st ls', 'gaussmf', [0.5 -4]); Wrong: fis = addmf(fis, 'input', 1, '1st ls', gaussmf, [0.5 -4]);