Date: Thu, 9 May 2024 | Total: 6
The manner epidemics occurs in a social network depends on various elements, with two of the most influential being the relationships among individuals in the population and the mechanism of transmission. In this paper, we assume that the social network has a homogeneous random topology of Erd\"os-R\'enyi type. Regarding the contagion process, we assume that the probability of infection is proportional to the proportion of infected neighbours. We consider a constant population, whose individuals are the nodes of the social network, formed by two variable subpopulations: Susceptible and Infected (SI model). We simulate the epidemics on this random network and study whether the average dynamics can be described using a mean field approach in terms of Differential Equations, employing the law of mass action. We show that a macroscopic description could be applied for low average connectivity, adjusting the value of the contagion rate in a precise function. This dependence is illustrated by calculating the transient times for each connectivity. This study contributes valuable insights into the interplay between network connectivity, contagion dynamics, and the applicability of mean-field approximations. The delineation of critical thresholds and the distinctive behaviour at lower connectivity enable a deeper understanding of epidemic dynamics.
We introduce the class of compactly H\"older mappings between metric spaces and determine the extent to which they distort the Minkowski dimension of a given set. These mappings are defined purely with metric notions and can be seen as a generalization of Sobolev mappings, without the requirement for a measure on the source space. In fact, we show that if $f:X\rightarrow Y$ is a continuous mapping lying in some super-critical Newtonian-Sobolev space $N^{1,p}(X,\mu)$, under standard assumptions on the metric measure space $(X,d,\mu)$, it is then a compactly H\"older mapping. The dimension distortion result we obtain is new even for Sobolev mappings between weighted Euclidean spaces and generalizes previous results of Kaufman and Bishop-Hakobyan-Williams.
We provide a sufficient and necessary condition in terms of the stoichiometric coefficients for a bi-reaction network to admit multistability. Also, this result completely characterizes the bi-reaction networks according to if they admit multistability.
For $C^{1+}$ maps, possibly non-invertible and with singularities, we prove that each homoclinic class of an adapted hyperbolic measure carries at most one adapted hyperbolic measure of maximal entropy. We then apply this to study the finiteness/uniqueness of such measures in several different settings: finite horizon dispersing billiards, codimension one partially hyperbolic endomorphisms with "large" entropy, robustly non-uniformly hyperbolic volume-preserving endomorphisms as in Andersson-Carrasco-Saghin, and strongly transitive non-uniformly expanding maps.
Impulsive dynamical systems, modeled by a continuous semiflow and an impulse function, may be discontinuous and may have non-intuitive topological properties, as the non-invariance of the non-wandering set or the non-existence of invariant probability measures. In this paper we study dynamical features of impulsive flows parameterized by the space of impulses. We prove that impulsive semiflows determined by a C1-Baire generic impulse are such that the set of hyperbolic periodic orbits is dense in the set of non-wandering points which meet the impulsive region. As a consequence, we provide sufficient conditions for the non-wandering set of a typical impulsive semiflow (except the discontinuity set) to be invariant. Several applications are given concerning impulsive semiflows obtained from billiard, Anosov and geometric Lorenz flows.
We prove that the canonical action of every hyperbolic group on its Gromov boundary has the shadowing (aka pseudo-orbit tracing) property. In particular, this recovers the results of Mann et al. that such actions are topologically stable.