"indexed assignment to a matrix element is not permitted in a solve block mathcad"

[jgvinholi]

Hi, I've been trying to use some solve blocks inside a program. The solve blocks return matrixes with the solutions of the equations. I'm trying to get these results inside a program, but the error "indexed assignment to a matrix element is not permitted in a solve block mathcad" keeps showing. I don't really understand what it means. I've attached the file.

Edit: I haven't put any solve block inside the program, but instead tried to get outputs from outside solve blocks.

[LucMeekes]

You cannot (in Mathcad or Prime) put a solve block into a program, but you can make the output of a solve block a function which you can use further.

Success!

Luc

[jgvinholi]

I haven't tried to put a solve block inside a program. What I'm doing is calling functions that give outputs from other solve blocks. What I'm doing is exactly what you suggested. I've expressed Myself badly, sorry.

Thanks for the response.

[LucMeekes]

Check your labels. The 'Hf' solve block (2nd iteration) tries to find 'Hf1[1' which is labelled as a constant (green) instead of variable (black).

Success!

Luc

[Werner_E · Accepted solution]

The reason for the error message you get is, that you have erratically used a vector index for H1_2 in the Hi-gap solve block when assigning the guess value.

After you change that for a literal index the error is gone away, but you get a new unit mismatch error when calling Hi_iron.

Can it be that you forgot a factor S3 on the left side of the third equation?

[jgvinholi]

Thank you very much. I'll take a look closely tomorrow and will post here a feedback.

[jgvinholi]

These really were the problems! In my ingenuity, I'd perhaps never get to this conclusion. Thank you.

[terryhendicott]

Hi,

In the last bit of programming for "ans" you have used an iteration of "i" from 0 to 50.

The index "i" is not used within the programming block?  Is this intentional?

Cheers

Terry

[Werner_E]

---

@terryhendicott wrote:

Hi,

In the last bit of programming for "ans" you have used an iteration of "i" from 0 to 50.

The index "i" is not used within the programming block?  Is this intentional?

Cheers

Terry

---

I guess this is just doing the iteration 50 times hoping for convergence.

I think that one single solve block can replace all the others and this loop a well:

[jgvinholi]

That's impressive! How does Mathcad gets Hf_1 to calculate Mu_f? I thought the only way I could solve this problem was through repetitions, as I had done. Does Mathcad "know" it is needed to make a iterative process and do it by itself? Thanks for the help.

[Werner_E]

---

@jgvinholi wrote:

That's impressive! How does Mathcad gets Hf_1 to calculate Mu_f? I thought the only way I could solve this problem was through repetitions, as I had done. Does Mathcad "know" it is needed to make a iterative process and do it by itself? Thanks for the help.

---

Yes, Mathcad "knows" that. Every numerical solve block uses a numeric iteration algorithm (something like Levenberg-Marquardt, Quasi-Newton, etc. - not sure what the algorithm in KNITRO used by prime is) to get its solutions. Also the solve blocks you implemented work that way.

I simply combined your two last solve blocks to let Mathcad play around with all seven variable and get a solution in one run.

Hope the factor S3 I added (as written in my first reply) to balance the units was the correct choice.

[jgvinholi]

Often, knowing how the tool you're using works makes a big difference. The S3 you added is absolutely correct. I had forgotten to add it.

Again I'd like to thank you for the help. Couldn't be more grateful of your response than I do.

[jgvinholi]

As Werner_E said, I'm just hopping for convergence heheh.