The second paper of January 2025 in Computo, by Daphné Giorgi, Sarah Kaakai and Vincent Lemaire, introduces the R package IBMPopSim, which facilitates the simulation of the random evolution of heterogeneous populations using stochastic Individual-Based Models (IBMs).
The package relies on a unified mathematical framework, based on thinning of Poisson measures, for the simulation of IBMs where individuals are represented by their birth date, death date (+∞ if none), and a collection of features. It uses Rcpp for efficiency.
The paper introduces this mathematical framework, gives a detailed overview of the IBMPopSim package, and illustrates it on two use cases, one from actuarial sciences and the other from population genetics.
On this last example, the authors show that their randomized algorithm is one or two orders of magnitudes faster than the full algorithm.
The paper is available with (of course) R code at: https://doi.org/10.57750/sfxn-1t05
#populationDynamics #openScience #openAccess #openSource #Rstats