# splitfngr v0.1.2

0

0th

Percentile

## Combined Evaluation and Split Access of Functions

Some R functions, such as optim(), require a function its gradient passed as separate arguments. When these are expensive to calculate it may be much faster to calculate the function (fn) and gradient (gr) together since they often share many calculations (chain rule). This package allows the user to pass in a single function that returns both the function and gradient, then splits (hence 'splitfngr') them so the results can be accessed separately. The functions provided allow this to be done with any number of functions/values, not just for functions and gradients.

# splitfngr

Sometimes multiple values are returned by a single function. This makes sense when their combined evaluation is faster than separate evaluations, such as function and corresponding gradient values which require solving the same matrix system. If the user wants to have separate functions for the function and gradient, such as when passing to an optimization routine.

Functions provided in this package allow this to be done in a general situation.

Below is an example. The R function optim with method BFGS requires a function for the function value and gradient value separately. In this trivial example the gradient value is not recalculated when it is called.

# A function that returns
quad_share <- function(x){list(sum(x^4), 4*x^3)}
# Run BFGS using function with gradient
splitfngr::optim_share(par=c(3, -5), quad_share, method="BFGS")
#> $par #> [1] -0.0003599207 0.0006973696 #> #>$value
#> [1] 2.532927e-13
#>
#> $counts #> function gradient #> 42 36 #> #>$convergence
#> [1] 0
#>
#> \$message
#> NULL


See the documentation for more examples and details on usage.

## Functions in splitfngr

 Name Description lbfgs_share Use splitfngr with lbfgs make_share Convert a function from multiple function arguments to a single function fngr Access a list of values separately but calculate them together. This function generalizes grad_share for any number of functions. nlminb_share Use splitfngr with nlminb optim_share Use splitfngr with optim grad_share Split function and gradient calculation No Results!