• View
• Edit
• History
• Print

PositiveParameters

< Mujpy.SaveFit | Index | Mujpy.Fitplot >

List of positive parity parameters that can be unconstrained, but one wants to converge to the positive value

• σ in mg, bg, jg
• Δ in kg, kd, ks

List of parameters that must be positive defined

• Λ,β in bs, ms, js, ks

Positive defined parameters must have fixed limits at all time. This is ensured by __available_components__ in mujpy.aux.aux invooked by the int2min methods

A positive parity parameter, like σ in a Gaussian relaxation, is defined as the parameter for which the function has positive parity: {$\exp\left[-\frac{\sigma^2t^2}2\right] = \exp\left[-\frac {(-\sigma)^2t^2}2\right]$}

The strategy here is to minimize twice, first fixing the parameter limits to [0,None] and the second time releasing them.

This is accomplished respectively by __available_components__ invoked by the int2min methods, and by positive_parity, all in mujpy.aux.aux

< Mujpy.SaveFit | Index | Mujpy.Fitplot >