Nlp

Nlp

Namespace

The JXG.Math.Nlp namespace holds numerical algorithms for non-linear optimization, mainly the port of COBYLA.

Methods

Own

(static) FindMinimum(calcfc, n, m, x, rhobeg, rhoend, iprint, maxfun, testForRoundingErrorsopt) → {Number}

Minimizes the objective function F with respect to a set of inequality constraints CON, and returns the optimal variable array. F and CON may be non-linear, and should preferably be smooth. Calls JXG.Math.Nlp#cobylb.

Parameters

Name	Attributes	Default	Description
`calcfc`			Interface implementation for calculating objective function and constraints.
`n`			Number of variables.
`m`			Number of constraints.
`x`			On input initial values of the variables (zero-based array). On output optimal values of the variables obtained in the COBYLA minimization.
`rhobeg`			Initial size of the simplex.
`rhoend`			Final value of the simplex.
`iprint`			Print level, 0 <= iprint <= 3, where 0 provides no output and 3 provides full output to the console.
`maxfun`			Maximum number of function evaluations before terminating.
`testForRoundingErrors`	<optional>	`false`

Returns

Exit status of the COBYLA2 optimization.

Type: Number

Details

Source: math/nlp.js, line 143

(static) cobylb(calcfc, n, m, mpp, x, rhobeg, rhoend, iprint, maxfun, testForRoundingErrorsopt) → {Number}

JavaScript implementation of the non-linear optimization method COBYLA.

Parameters

Name	Type	Attributes	Default
`calcfc`	function
`n`	Number
`m`	Number
`mpp`	Number
`x`	Number
`rhobeg`	Number
`rhoend`	Number
`iprint`	Number
`maxfun`	Number
`testForRoundingErrors`	Boolean	<optional>	`false`

Returns

Exit status of the COBYLA2 optimization

Type: Number

Details

Source: math/nlp.js, line 248

Inherited

none

Details

Source: math/nlp.js, line 85