different Opti-method makes different paramter-results but same SSE?

Hi Friends,

one more question.

Here my sheets: In the first area I used Quasi-Newton and got fine results, nearly the same as the ideal values.

But when I use LM I get much other results but both methods have the same SSE. could this be true?

Thanks

Author

12-Amethyst

Here the sheet.

Translate

25-Diamond I

> could this be true?

I guess yes. An optimization process can have multiple solutions of nearly equal "quality".

It depends how your problem is conditioned, what the error surface looks like. Generally hard to say.

Translate

3 replies