mosek_MIPsearch {DoE.MIParray} R Documentation

## Functions to Search for optimum MIP Based Array Using Gurobi or Mosek

### Description

The functions search through different orderings of the nlevels vector with the goal to create an array with minimum resolution and optimized shortest word length. They create the orders and call gurobi_MIParray or mosek_MIParray for each order.

### Usage

mosek_MIPsearch(nruns, nlevels, resolution = 3, maxtime = 60,
stopearly=TRUE, listout=FALSE, orders=NULL,
distinct = TRUE, detailed = 0, forced=NULL, find.only=TRUE,
nthread=2, mosek.opts = list(verbose = 1, soldetail = 1),
mosek.params = list(dparam = list(LOWER_OBJ_CUT = 0.5, MIO_TOL_ABS_GAP = 0.2,
INTPNT_CO_TOL_PFEAS = 1e-05, INTPNT_CO_TOL_INFEAS = 1e-07),
iparam = list(PRESOLVE_LINDEP_USE="OFF", LOG_MIO_FREQ=100)))
gurobi_MIPsearch(nruns, nlevels, resolution = 3, maxtime = 60,
stopearly=TRUE, listout=FALSE, orders=NULL,
distinct = TRUE, detailed = 0, forced=NULL, find.only=TRUE,
nthread = 2, heurist=0.5, MIQCPMethod=0, MIPFocus=1,
gurobi.params = list(BestObjStop = 0.5, OutputFlag=0))


### Value

an array of class oa with the attributes added by mosek_MIParray or gurobi_MIParray, resp.
In addition, the attribute optorder contains the vector of level orders that yielded the best design; if listout=TRUE, also the attributes orders and allplans.

Objects with the attribute allplans are quite large. If the attribute is no longer needed, it can be removed from an object named obj (replace with the name of your object) by the command
attr(obj, "allplans") <- NULL

Ulrike Groemping

### References

Groemping, U. and Fontana R. (2019). An Algorithm for Generating Good Mixed Level Factorial Designs. Computational Statistics & Data Analysis 137, 101-114.

Groemping, U. (2020). DoE.MIParray: an R package for algorithmic creation of orthogonal arrays. Journal of Open Research Software, 8: 24. DOI: https://doi.org/10.5334/jors.286

### See Also

See also mosek_MIParray and gurobi_MIParray, oa_feasible from package DoE.base for checking feasibility of requested array strength (resolution - 1) for combinations of nruns and nlevels, and lowerbound_AR from package DoE.base for a lower bound for the length R words in a resolution R array.

### Examples

## Not run:
## can also be run with gurobi_MIParray instead of mosek_MIParray
## there are of course better ways to obtain good arrays for these parameters
## (e.g. function FrF2 from package FrF2)
oa_feasible(18, c(2,3,3,3,3), 2)  ## strength 2 array feasible
lowerbound_AR(18, c(2,3,3,3,3), 3)  ## lower bound for A3
## of course not necessary here, the design is found fast
feld <- mosek_MIPsearch(18, c(2,3,3,3,3), stopearly=FALSE, listout=TRUE, maxtime=30)
## even stopearly=TRUE would not stop, because the lower bound 2 is not achievable
feld
names(attributes(feld))
attr(feld, "optorder")
## even for this simple case, running optimization until confirmed optimality
## would be very slow

## End(Not run)


[Package DoE.MIParray version 1.0 Index]