[R-sig-dyn-mod] Question about the deSolve package

[Yichi Zhang]

To whom it may concern,

I have a question regarding the integration rationale behind the solvers in
the package. That is, if my system evolves with time t, does the simulated
stage (like infections in the susceptible-infection model) at a certain
time point uses information at later time points? I had this question
because I forced an interpolating function of the observed intervention
trajectory C(t) that modifies the SI system. It appeared that to simulate
infections until t=40 or t=20 will affect my simulated infections when t is
in (0, 20). Does this suggest simulated infections borrow information on
the interpolated intervention function at a later time? Will a structural
model that stipulates C(t) is decided by previous infections refrain from
using info at later times?

Thank you for your time and consideration!

Best regards,
Yichi Zhang

-- 
Yichi Zhang
PhD Student in Biostatistics
Yale School of Public Health

	[[alternative HTML version deleted]]