Ctrl+K

Simplex#

See Nelder-Mead Simplex for more information on the Nelder-Mead Simplex method.

minimisers.simplex.fMinSearch(funfcn, x, options, dis, varargin)#

Perform a Nelder-Mead Simplex minimisation of an objective function.

Parameters:

  • funfcn (function) – The function to minimise.

  • x (vector) – The starting point to minimise from.

  • options (struct) – Controls options for the algorithm.

  • dis (string) – Options for how to display algorithm output.

  • varargin (options) – Other options from the Mathworks fminsearch.

Returns:

  • x (vector) – The minimum value found.

  • fval (vector) – The objective function value at output point x.

  • exitflag (integer) –

    Reason fMinSearch stopped:

    • 1 if converged;

    • 0 if max iterations/evals exceeded;

    • -1 if the output function returned an error.

  • output (struct) – Diagnostic information about the optimisation process.

Notes

See also MathWorks fminsearch doc page: https://www.mathworks.com/help/matlab/ref/fminsearch.html

minimisers.simplex.runSimplex(problemStruct, controls)#

Run the Nelder-Mead simplex algorithm for a given problem and controls.

Parameters:

  • problemStruct (struct) – the Project struct.

  • controls (struct) – the Controls struct.

Returns:

  • problemStruct (struct) – the output project struct.

  • result (struct) – the calculation and optimisation results object.

minimisers.simplex.simplexIntrafun(x, problemStruct, controls, params)#

Produce a chi-squared fit value at a point for a given problem and controls.

Used as the objective function for Nelder-Mead simplex.

Parameters:

  • x (vector) – The point to evaluate.

  • problemStruct (struct) – The problem struct for calculation.

  • controls (struct) – The controls struct for the calculation.

  • params (struct) – A struct containing parameter fit information.

Returns:

  • fval (float) – The chi-squared value at x.

  • result (struct) – The result struct from the reflectivity calculation.

minimisers.simplex.simplexXTransform(x, params)#

Convert unconstrained variables into their original domains.

Parameters:

  • x (vector) – The point to transform.

  • params (struct) – Parameter fit information.

Returns:

xtrans – The point x transformed for the parameter bounds.

Return type:

vector