Binding and unbinding events in molecular systems happen on timescales that brute-force molecular dynamics can't reach. A potassium ion entering a crown ether takes microseconds to milliseconds; the simulation timestep is femtoseconds. The gap is six to nine orders of magnitude.
The authors (arXiv:2601.09187) introduce IEPDYN — an integral-equation formalism for population dynamics. Instead of running one long trajectory and waiting for the rare event, they define configurational states along the reaction coordinate and track how probability density flows between neighboring states. The Liouville equation governs these flows. A Markov approximation converts the problem into integral equations that can be solved from short simulations of each state.
The key advantage over competing methods (Markov state models, milestoning): no lag-time dependence. MSMs require choosing a lag time — too short and the model isn't Markov, too long and you lose temporal resolution. The choice is often more art than science and affects the results. IEPDYN formulates the dynamics as integral equations that avoid this arbitrary parameter.
For the crown ether–potassium system, IEPDYN matches brute-force simulations while reducing computational time by approximately two orders of magnitude.
The through-claim: the rare-event problem isn't a timescale problem — it's a formulation problem. The dynamics are already Markovian at the level of probability currents between states. The integral-equation approach makes this explicit without requiring the user to guess an appropriate discretization timescale.