Spine for interacting populations and sampling

Vincent Bansaye (1, 2)

1 CMAP - Centre de Mathématiques Appliquées - Ecole Polytechnique
2 MERGE - Mathématiques pour l’évolution, la reproduction, la croissance et l’émergence

We consider Markov jump processes describing structured populations with interactions via density dependance. We propose a Markov construction with a distinguished individual which allows to describe the random tree and random sample at a given time via a change of probability. This spine construction involves the extension of type space of individuals to include the state of the population. The jump rates outside the spine are also modified. We apply this approach to some issues concerning evolution of populations and competition. For single type populations, we derive the diagram phase of a growth fragmentation model with competition and the growth of the size of birth and death processes with multiple births. We also describe the ancestral lineages of a uniform sample in multitype populations.

Mots clés: Jump Markov process; random tree; spine; interactions; positive semigroup; martingales; population models; Jump Markov process

Domaines: Probabilités [math.PR]

Vincent Bansaye. Spine for interacting populations and sampling. Bernoulli, 2024, 30 (2), ⟨10.3150/23-BEJ1645⟩. ⟨hal-03219097v3⟩