Common output for all CGO solvers

From TomWiki
Jump to navigationJump to search

Output parameters

Result structure

Structure with results from optimization

The following fields are used
Field Description
x_k Matrix with the best points as columns, f(x_k) == f_k
Iter Number of iterations
FuncEv Number of function evaluations
ExitText Text string giving ExitFlag and Inform information
ExitFlag Always 0, except

1 = Initial interpolation failed, normally because too huge f(x) (rbfSolve only)

Inform Value indicating the type of termination. See the table for the specific solver

Save file

To make a warm start possible, all CGO solvers save the following information to the file cgoSave.mat

The following variables are saved
Name Name of the problem
WSInfo Structure with warmstart information, see table below

If the cgoSave.mat file fails to open for writing, the information is also available in Result.CGO.WarmStartInfo. Through a call to WarmDefGLOBAL, the Prob structure can be setup for warm start. In this case, the solver will not load the data from cgoSave.mat.

WSInfo

Structure with warm start information

The following fields are used
Name Name of the problem
OMIN Matrix d x nLocal with nLocal minima, in original coordinates, found during last or previous runs.
FMIN Column vector with nLocal costly function values f(x) computed at the nLocal points in OMIN
nInit Number of initial points >= d+1 (2d if center points)
O Matrix with sampled points (in original space)
X Matrix with sampled points (in unit space if Prob.CGO.SCALE == 1)
F00 Vector of pure function values, before penalties.
Cc Matrix with COSTLY costraint values, Cc(x).

Non-costly constraint values, Cn(x), are always recomputed using the linear constraints and noncostly nonlinear c(x).

F Vector with function values (penalty added for costly Cc(x))
Fpen Vector with function values + additional penalty if infeasible
Feasible If Feasible = 1, at least one feasible point has been found
fMin Lowest function values found
CcMin Costly constraint values at best point found. Now empty.
fMinIter Iteration when (xMin, fMin) was found
fMinIdx Iteration when (xMin, fMin) was found
rngState Current state of the random number generator used