Optimization

What Optimization does

Fitting study

Minimize a function or fit PHREEQC parameters to measured data with Ceres Solver and its trust-region methods, Levenberg–Marquardt or dogleg.

Any parameter, any target

Fit rate constants, sorption parameters or exchange capacities against any column of your measured data, scalar or vector.

Control over the search

Bounds keep the fit physically sensible, a parameter can be held fixed or searched on a log scale, and residuals of different reliability can be weighted by their standard deviation.

The fit is a value

The study keeps the target run at the best parameters found, and each fitted value can be used in an expression downstream rather than copied by hand.

Typical uses

• Fitting kinetic rate constants
• Calibrating Langmuir or surface complexation parameters
• Matching titration curves
• Fitting a rate constant that spans orders of magnitude, on a log scale