search

tgvaughan / MASTER

A versatile simulation engine for stochastic population dynamics models.
http://tgvaughan.github.io/MASTER
GNU General Public License v3.0
22 stars 8 forks source link

Finite time step trajectory probability calculations #37

Closed tgvaughan closed 9 years ago

tgvaughan commented 10 years ago

A partial implementation of a method for computing trajectory probabilities is now in the master branch. This seems correct for Gillespie trajectories (both inheritance and not) but gives quite different results when using finite timestep integration.

The reason behind this is that finite time-step trajectories yield actual probabilities, while the Gillespie trajectories yield densities. I think the finite time step results should be brought into agreement with the Gillespie results. This can be done by including the missing delta t terms, the size of which depend on the step size chosen and the number of which depend on the number of events included in the trajectory.

tgvaughan commented 10 years ago

Commit 237949d020d0fc32ba7770fa94627701489845cd brings the different calculations into line.

tgvaughan commented 9 years ago

Closed by 237949d.