Using solution of DDE model as history for another (larger) DDE problem

I want to study effects on a DDE model when I add another component to the system.

Using resize! and callbacks doesn’t work.

Therefore I want to solve the DDE model for n components across some time span, then use this as the initial condition and system history for a second DDE problem with n + 1 components. (The history of the n+1 component would be also defined here, probably something constant).

However it is not clear to me how to use the output of solve(prob, alg) to define the history function h(p, t) to achieve this, as solve returns a vector with (often) variable timesteps between values.

From reading the DifferentialEquations.jl docs, it looks like h is assumed to be a continuous function, so presumably the internal mechanisms in solve create some continuous interpolation between solved values as a proxy for h? If this is true I’d like to replicate this as close as possible.

Does anyone have any suggestions?

The output of solve is a sol(t) function so h(p,t) = sol(t) just works.

1 Like