errorLocalizer.mip: Localize errors using a MIP approach.
Description
Localize errors using a MIP approach.Usage
errorLocalizer.mip(E, x, weight = rep(1, length(x)),
maxduration = 600, verbose = "neutral", ...)Arguments
weight
numeric with weights
maxduration
number of seconds that is spent on
finding a solution
verbose
verbosity argument that will be passed on
to solve lpSolveAPI
...
other arguments that will be passed on to
solve.
Value
- list with solution weight
w, logical
adapt stating what to adapt, x_feasible and
the lp problem (an lpExtPtr object)
Details
errorLocalizer.mip uses E and x to
define a mixed integer problem and solves this problem
using lpSolveApi. This function can be much faster
then errorLocalizer but does not return the
degeneracy of a solution. However it does return an
bonus: x_feasible, a feasible solution.Note
If the maximim absolute value of $x\geq 10^9$ 1E9, it
is pre-scaled with a factor of $\sqrt{\max(|x|)}$.References
E. De Jonge and Van der Loo, M. (2012) Error localization
as a mixed-integer program in editrules (included with
the package)
lp_solve and Kjell Konis. (2011). lpSolveAPI: R Interface
for lp_solve version 5.5.2.0. R package version
5.5.2.0-5. http://CRAN.R-project.org/package=lpSolveAPI