R help - Apr 2011 - Fwd: The results of your email commands

Hi R community,

I posted a question on using the R maximum likelihood functions a short
while ago and got an email saying that some of the content was
"unprocessed". Hence, I am reposting the question just to be sure
(sorry for
the multiple emails if both reached you).

My question is reagarding the way we use constraints in programs like
maxBFGS, etc.

We need to put two matrices into a list, containing the mutipliers on the
LHS of the constraint equations and the constants on the RHS.
I want to constrain the last parameter of my routine to a negative value.

Hence, I specify:
lhs=matrix(0,nrow=noOfParameters,ncol=noOfParameters)
lhs[noOfParameters,noOfParameters]=-1

rhs=matrix(0,nrow=(noOfParameters))

Then I use:
constraints = list(ineqA=lhs,ineqB=rhs)
in the maxBFGS function.

However, the result is that no matter what parameters I enter (positive,
negative or 0), I get an error saying that the initial value is out of
bounds.

Can you please tell me what I am doing wrong?

Thanks in advance,

Rohit

	[[alternative HTML version deleted]]