optmatch (version 0.7-5)

minControlsCap: Set thinning and thickening caps for full matching

Description

Functions to find the largest value of min.controls, or the smallest value of max.controls, for which a full matching problem is feasible. These are determined by constraints embedded in the matching problem's distance matrix.

Usage

minControlsCap(distance, max.controls = NULL, subclass.indices = NULL)
maxControlsCap(distance, min.controls = NULL, subclass.indices = NULL)

Arguments

distance

Either a matrix of nonegative, numeric discrepancies, or a list of such matrices. (See fullmatch for details.)

max.controls

Optionally, set limits on the maximum number of controls per matched set. (Only makes sense for minControlsCap.)

min.controls

Optionally, set limits on the minimum number of controls per matched set. (Only makes sense for maxControlsCap.)

subclass.indices

This argument no longer supported (nor necessary).

Value

• For minControlsCap, strictest.feasible.min.controls and given.max.controls. For maxControlsCap, given.min.controls and strictest.feasible.max.controls.
• strictest.feasible.min.controls
  The largest values of the fullmatch argument min.controls that yield a full match;
• given.max.controls
  The max.controls argument given to minControlsCap or, if none was given, a vector of Infs.
• given.min.controls
  The min.controls argument given to maxControlsCap or, if none was given, a vector of 0s;
• strictest.feasible.max.controls
  The smallest values of the fullmatch argument max.controls that yield a full match.

Details

The function works by repeated application of full matching, so on large problems it can be time-consuming.

References

Hansen, B.B. and S. Olsen Klopfer (2006), Optimal full matching and related designs via network flows, Journal of Computational and Graphical Statistics 15, 609--627.

See Also

fullmatch

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.
plantdist <- matrix(nrow=7, ncol=19,byrow=TRUE,data=c(
28, 0, 3,22,14,30,17,28,26,28,20,22,23,26,21,18,34,40,28,
24, 3, 0,22,10,27,14,26,24,24,16,19,20,23,18,16,31,37,25,
10,18,14,18, 4,12, 6,11, 9,10,14,12, 6,14,22,10,16,22,28,
 7,28,24, 8,14, 2,10, 6,12, 0,24,22, 4,24,32,20,18,16,38,
17,20,16,32,18,26,20,18,12,24, 0, 2,20, 6, 8, 4,14,20,14,
20,31,28,35,20,29,22,20,14,26,12, 9,22, 5,15,12, 9,11,12,
14,32,29,30,18,24,17,16,10,22,12,10,17, 6,16,14, 4, 8,17),
dimnames=list(c("A","B","C","D","E","F","G"),
c("H","I","J","K","L","M","N","O","P","Q","R",
"S","T","U","V","W","X","Y","Z")))

(tmn <- minControlsCap(plantdist)$strictest)
maxControlsCap(plantdist, min=tmn)
splitdist <- list(one=plantdist[1:3, 1:9], two=plantdist[4:7, 10:19])
(tmn <- minControlsCap(splitdist)$strictest)
maxControlsCap(splitdist, min=tmn)