search for: nlmx

...ie: test.outer<-function(param.outer){ slope<-nlm(test.inner,param.inner) ... loglikelihood<-sum(...) return(-loglikelihood) } and nlm(test.outer,param.outer) on the second call of test.outer by nlm, the inner calls to test.inner are finding param.outer. I have tried the obvious nlmx<-nlm and calling nlmx as the inner call but this does not fix it. Can I use recursion with nlm or is there an alternative non-linear minimiser? I suspect that it is the static memory within the Fortran - is there any way of copying this to an object after the outer call then restoring it after...