get.solver.params {ManifoldOptim}

Get parameters to initialize solver

Description

Get parameters to initialize solver

Usage

get.solver.params(
  isconvex = FALSE,
  DEBUG = 0,
  Tolerance = 1e-04,
  Max_Iteration = 1000,
  IsCheckParams = FALSE,
  IsCheckGradHess = FALSE,
  ...
)

Arguments

isconvex

Indicator for whether the function is convex (TRUE or FALSE)

DEBUG

Verbosity level in {0,1,2,3}. Use 0 for quietest with no messages printed. Use 3 for most verbose,

Tolerance

Tolerance used to assess convergence. See Huang et al (2016b) for details on how this is used,

Max_Iteration

Maximum iterations to be used by the solver (a non-negative integer),

IsCheckParams

Should solver check inputs and print summary message before optimization (TRUE or FALSE),

IsCheckGradHess

Check correctness of the gradient and Hessian functions (TRUE or FALSE).

...

Additional arguments to pass to the solver. These are not validated by the get.solver.params function. Users should refer to the C++ library's user manual for available arguments.

Details

Solver-specific parameters may also be added to the object returned from get.solver.params, via standard list manipulation. Interested users should refer to Huang et al (2016b) for available options.

Value

List containing input arguments for solver

References

Wen Huang, P.A. Absil, K.A. Gallivan, Paul Hand (2016a). "ROPTLIB: an object-oriented C++ library for optimization on Riemannian manifolds." Technical Report FSU16-14, Florida State University.

Wen Huang, Kyle A. Gallivan, and P.A. Absil (2016b). Riemannian Manifold Optimization Library. URL https://www.math.fsu.edu/~whuang2/pdf/USER_MANUAL_for_2016-04-29.pdf

S. Martin, A. Raim, W. Huang, and K. Adragni (2020). "ManifoldOptim: An R Interface to the ROPTLIB Library for Riemannian Manifold Optimization." Journal of Statistical Software, 93(1):1-32.