Confidence intervals (granted)

model <- mxModel(....)
modelOut <- mxRun(model)
library(snowfall)
sfInit(parallel = TRUE, cpus = ???)
sfLibrary(OpenMx)
intervals1 <- mxModel(modelOut, mxCI('a'), name = "intervals1", independent = TRUE)
intervals2 <- mxModel(modelOut, mxCI('b'), name = "intervals2", independent = TRUE)
intervals3 <- mxModel(modelOut, mxCI('c'), name = "intervals3", independent = TRUE)
intervals4 <- mxModel(modelOut, mxCI('d'), name = "intervals4", independent = TRUE)
intervals5 <- mxModel(modelOut, mxCI('e'), name = "intervals5", independent = TRUE)
topModel <- mxModel('topModel', intervals1, intervals2, intervals3, intervals4, intervals5)
results <- mxRun(topModel, intervals = TRUE)
sfStop()

UPDATE: I forgot to give the submodels new names. The name collision would have resulted in only one submodel stored as a child of the topModel.

modelNameAddPrefix <- function(model, prefix) {
   if (missing(model) || !is(model, "MxModel")) {
      stop("first argument must be a MxModel object")
   }
   if (missing(prefix) || length(prefix) != 1 || !is.character(prefix)) {
      stop("second argument must be a character string")
   }
   model <- modelNameAddPrefixHelper(model, model, prefix) 
   return(model)
}

modelNameAddPrefixHelper <- function(job, model, prefix) {
   job <- mxRename(job, paste(prefix, model@name, sep = ''), model@name)
   if (length(model@submodels) > 0) {
      for(i in 1:length(model@submodels))  {
          job <- modelNameAddPrefixHelper(job, model@submodels[[i]], prefix)
      }
   }
   return(job)
}

So when I tested it using foo1 <- modelNameAddPrefix(twinACEModel, "foo1_") the resulting model is named 'foo1_twinACE' and has three submodels 'foo1_common', 'foo1_MZ', 'foo1_DZ'. It's not exactly user-friendly at the moment, but it will help in running the confidence intervals.

Oh, plus I found a bug in the function mxRename(). It will currently not rename the appropriate confidence intervals, because confidence intervals had not been implemented when mxRename() was written. But you should be specifying the confidence intervals manually after creating all the submodels for this particular workflow.

UPDATE: Oops, I forgot that child models could refer to entities in the parent models. I had to change the recursion step to apply to the entire nested model. The code has been updated.

> require(OpenMx)
> data(demoOneFactor)
> factorModel <- mxModel("One Factor",
+ mxMatrix("Full", 5, 1, values=0.2,free=T, name="A"),
+ mxMatrix("Symm", 1, 1, values=1,free=F, name="L"),
+ mxMatrix("Diag", 5, 5, values=1,free=T, name="U"),
+ mxAlgebra(A %*% L %*% t(A) + U, name="R"),
+ mxCI(c('A[1,1]','R[1,1]')),
+ mxMLObjective("R", dimnames = names(demoOneFactor)),
+ mxData(cov(demoOneFactor), type="cov", numObs=500))
> factorModelOuput <- (mxRun(factorModel,intervals=T))
Running One Factor
Warning messages:
1: In model 'One Factor' while computing the lower bound for 'One Factor.A[1,1]' NPSOL returned a non-zero status code 6. The model does not satisfy the first-order optimality conditions to the required accuracy, and no improved point for the merit function could be found during the final linesearch (Mx status RED)
2: In model 'One Factor' while computing the lower bound for 'One Factor.R[1,1]' NPSOL returned a non-zero status code 6. The model does not satisfy the first-order optimality conditions to the required accuracy, and no improved point for the merit function could be found during the final linesearch (Mx status RED)
3: In model 'One Factor' while computing the upper bound for 'One Factor.A[1,1]' NPSOL returned a non-zero status code 6. The model does not satisfy the first-order optimality conditions to the required accuracy, and no improved point for the merit function could be found during the final linesearch (Mx status RED)
4: In model 'One Factor' while computing the upper bound for 'One Factor.R[1,1]' NPSOL returned a non-zero status code 6. The model does not satisfy the first-order optimality conditions to the required accuracy, and no improved point for the merit function could be found during the final linesearch (Mx status RED)
> factorModelSummary <- summary(factorModelOutput)
> factorModelOutput@output$confidenceIntervals
[,1] [,2]
>

And that's it. Summary is no help either.