Theoretical background on constraint handling

[Ahmed Abouhussein]

Hello and happy new years!

I wanted to ask about any references on the theory behind constraint handling. How does lmft-py actually handle bound, inequality and equality constraints?

I looked here, through the release notes and the google group but could not find any information on the theoretical basis, only on the application. 

Any information is appreciated.

[Matt Newville]

Hi Ahmed,

The bounds mechanism is described at
https://lmfit.github.io/lmfit-py/bounds.html which includes some
references, say the MINUIT documentation. The theory is to map
unbounded values to bounded values in a way that is reversible and
smooth.

Constraints in lmfit are done by setting one parameter's value to be
an algebraic expression of the other parameter values and maybe a set
of known constants and functions. A reference for that is either
"algebra" or the lmfit documentation/code.

Cheers,
--Matt