Grups de recerca
11
abr
Seminari FMC: Lucía Ramirez
El proper 11 d'abril podrem sentir la Dra. Lucía Ramirez, postdoc al grup de la Dra. M. A. Serrano. Us esperem a totes.
Dates:
11-04-2024
Horari:
12:45
Lloc:
Seminar room 320, 3rd floor of the Faculty of Physics, UB
Afegeix-ho a l'agenda (iCal)
ORDERING DYNAMICS AND PATH TO CONSENSUS IN MULTISTATE VOTER MODELS
Lucía Ramirez (luciaramirez@ub.eu)
Departament de Física de la Matèria Condensada, Universitat de Barcelona.
One of the most popular classes of models of opinion dynamics is the so-called voter model (VM). VMs describe populations of individuals located in the nodes of an interacting network, each characterized by a typically binary spin variable called opinion. The principal mechanism of change is imitation i.e., one individual copies the opinion state of another individual following a simple rule. Many variants of the VM have been introduced to capture different features of social interaction, including the Multistate Voter Model (MSVM). Unlike the traditional VM, the MSVM acknowledges that agents can hold one of multiple opinion states. Here, I will present results on the study of the MSVM with a focus on the path to consensus and ordering dynamics.
To do so, the time evolution of the density of active links is analyzed both at individual realizations and ensemble averages levels. The dynamics span all-to-all interaction and uncorrelated networks. Simple assumptions are derived to analytically calculate the average density of active links, finding an exponential decay with a timescale set by the size and geometry of the interaction network. Individual realizations reveal a sequence of opinion eliminations leading to consensus, with the population remaining in metastable states between eliminations, whose characteristics are studied in detail.
Furthermore, results on the ordering dynamics of nonlinear voter models are also presented. The ordering process varies based on whether the dynamics favor majority or minority opinions, and unlike linear models, the nonlinear model cannot be reduced to an effective two-state model. The analysis also presents a pair approximation for MSVM on networks, enhancing previous approximations and providing insights into the dynamics of multistate systems.
