[R] curve peeling

[Greg Tarpinian]

R2.3.0, WinXP.

I am trying to fit the model

   Y(t) = C1*exp(-exp(a)*t) + C2*exp(-exp(b)*t) - (C1+C2)*exp(-exp(c)*t)

to some pharmacokinetic data.  The data derives from a single oral dose
of a drug.  Does anyone know how to use peeling and/or other methods to 
obtain good starting estimates of the parameters in the triexponential?

The book "Mixed Effects Models in S and SPLUS" describes self-starting 
models for the biexponential and first-order one-compartment PK models,
but not for the triexponential.


Much thanks in advance,

   Greg