prDiGGer {DiGGer} | R Documentation |
Generate a DiGGer search for designs with some unreplicated treatments.
prDiGGer(numberOfTreatments, rowsInDesign, columnsInDesign, blockSequence = NULL, betweenRowCorrModel = "AR", betweenRowCorr = 0.5, betweenColumnCorrModel = "AR", betweenColumnnCorr = 0.5, treatName = NULL, treatNumber = NULL, treatRepPerRep = NULL, treatGroup = NULL, maxInterchanges = 20000, targetGroup = 1, runSearch = FALSE, splitOpt = NULL, rngSeeds = NULL)
numberOfTreatments |
Number of distinct treatments in the design. |
rowsInDesign |
Number of rows in the design. |
columnsInDesign |
Number of columns in the design. |
blockSequence |
List of dimension pairs of blocks to be optimised in sequence. |
betweenRowCorrModel |
The correlation pattern between rows may be "AR" AutoRegressive, "MA" Moving Average, or "ID" no correlation. |
betweenRowCorr |
Valid parameter values p, for "AR" -1<p<1, for "MA" -0.5<p<0.5 and for "ID" p=0. |
betweenColumnCorrModel |
As for |
betweenColumnCorr |
As for |
treatName |
Vector of texts giving the names of the treatment. |
treatNumber |
Vector of numbers associated with the treatments. |
treatRepPerRep |
Vector of replication levels for each treatment over the whole design. |
treatGroup |
Vector of group codes (up to 200 distinct values) associated with treatments. Group codes may be used to modify the A-efficiency measure in the optimisation. |
maxInterchanges |
Maximum number of interchanges used in the
search to spatially optimise the block design created by the call to
|
targetGroup |
Test treatment group used for optimisation in the spatial search phase. |
runSearch |
Logical value, whether to run the final spatial search immediately. |
splitOpt |
Dimension pair used for sub-optimal spatial distribution of replicated entries. |
rngSeeds |
Seeds c(s1,s2) to control the DiGGer search. s1 must be in the range [0,31328], s2 must be in the range [0,30081]. |
prDiGGer
creates an optimised block design for the replicated
treatments using the blocking sequence specified by the
blockSequence
parameters Each dimension pair is used in a
search within DiGGer. The optimised block
design for replicated treatments is augmented with the unreplicated
treatments to give an
initialDesign
of a DiGGer call. The DiGGer
object
created has the specifications for a search, using the spatial
parameters in the prDiGGer
call together with row and column
blocks.
rowColumn=TRUE
, to spatially optimise treatments within the
final blocks. The run
method must be used to search for the
optimised design. For targetGroup=1
the optimisation uses
aType="A11"
to target comparisons between Group 1 treatments of
interest.
A DiGGer
object with the specifications for a search or the
results of a DiGGer
search.
Neil Coombes
Coombes, N.E. (2002) The Reactive Tabu Search for Efficient Correlated Experimental Designs. PhD Thesis, Liverpool John Moores University.
Cullis, B.R., Smith, A.B. and Coombes, N.E. (2006) On the design of early generation variety trials with correlated data. JABES 11, 381–393.
Herzberg, A.M. and Jarrett, R.G. (2007). A-optimal block designs with additional singly replicated treatments. Journal of Applied Statistics 34, 61–70.
# design for 184 entries [23x10] with block sequence # [23x5], [6x5] and [6x1]. # The first 179 treatments are of interest - Group 1 prep184 <- prDiGGer(numberOfTreatments=184, rowsInDesign=23, columnsInDesign=10, treatRepPerRep=rep(c(2,1,5),c(26,153,5)), treatGroup=rep(c(1,2),c(179,5)), blockSequence=list(c(23,5),c(6,5),c(6,1))) prep184 # run the search to spatially optimise the design prep184 <- run(prep184) # get the matrix and plot it with checks in rep # duplicated treatments of interest in yellow. m184 <- getDesign(prep184) desPlot(m184,seq(153)+26,col=8,new=TRUE,label=FALSE) desPlot(m184,seq(26),col=7,new=FALSE,label=TRUE) desPlot(m184,seq(5)+179,col=2,new=FALSE,label=TRUE, bdef=cbind(6,1),bcol=4,bwd=4)