12.07.2015 Views

What Is Optimization Toolbox?

What Is Optimization Toolbox?

What Is Optimization Toolbox?

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

fsolveJacobPatternMaxPCGIterPrecondBandWidthTolPCGSparsity pattern of the Jacobian for finitedifferencing. If it is not convenient to computethe Jacobian matrix J in fun, lsqnonlin canapproximate J viasparsefinitedifferencesprovided the structure of J—i.e., locationsof the nonzeros—is supplied as the valuefor JacobPattern. In the worst case, if thestructure is unknown, you can set JacobPatternto be a dense matrix and a full finite-differenceapproximationiscomputedineachiteration(this is the default if JacobPattern is not set).This can be very expensive for large problems,so it is usually worth the effort to determine thesparsity structure.Maximum number of PCG (preconditionedconjugate gradient) iterations (see “Algorithm”on page 8-119).Upper bandwidth of preconditioner for PCG. Bydefault, diagonal preconditioning is used (upperbandwidth of 0). For some problems, increasingthe bandwidth reduces the number of PCGiterations. Setting PrecondBandWidth to 'Inf'uses a direct factorization (Cholesky) ratherthan the conjugate gradients (CG). The directfactorization is computationally more expensivethan CG, but produces a better quality steptowards the solution. The direct factorizationis computationally more expensive than CG,but produces a better quality step towards thesolution.Termination tolerance on the PCG iteration.Medium-Scale Algorithm OnlyThese options are used only by the medium-scale algorithm:8-115

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!