lbound = hard value?

boundbug.r

data(myFADataRaw, package = "OpenMx")
latents = c("G")
manifests = names(myFADataRaw)
myFADataRaw = myFADataRaw[, manifests]
m1 <- mxModel("m1", type="RAM", manifestVars = manifests, latentVars = latents,
	mxPath(from = latents, to = manifests, values = 0),
	mxPath(from = manifests, arrows = 2), # manifest residuals 
	mxPath(from = latents, arrows = 2, free = F, values = 1), # latents fixed@1
	mxData(cov(myFADataRaw, use = "complete"), type = "cov", numObs = nrow(myFADataRaw))
)
m1 = mxRun(m1)
summary(m1)$parameters[ ,c(1,5)]
m2 = mxModel(m1, name="bounded", mxPath(from = latents, to = manifests, values = 0, lbound  = .1))
m2 = mxRun(m2)
summary(m2)$parameters[ ,c(1,5)]

just CSOLNP ignore

This is a CSOLNP bug.

add

cbind(summary(m2)$parameters[ ,c(1,5)], summary(m1)$parameters[ ,c(1,5)])

to the end

And try re-running with

mxOption(NULL, "Default optimizer", "NPSOL")

CSOLNP apparently does not yet move parameters inside their bounds before starting.

Cheers,
Mike Hunter