Figure 4
From: Modeling fashion as an emergent collective behavior of bored individuals
Altering boredom in a homogenous population can critically affect collective trend dynamics: (A) Schematic of how changes in the boredom function of the model can be used to implement different degrees of boredom in the population. A right-shift (higher x0 parameter) of the boredom function simulates higher boredom susceptibility among agents, whereas a left-shift (lower x0 parameter) simulates a lower boredom susceptibility. (B) Manipulated boredom function, used to simulate low boredom proneness in the modeled population. (C) Equivalent manipulation to simulate increased boredom proneness. (D) Simulated trend color dynamics in a population with low boredom proneness (boredom function of B, ni = 200 individuals, memory size m = 12). Upper panel: Individual color choices over trials with converge to a single color. Middle panel: Trend color of the population over time. Lower panel: Uniformity of the population over the simulated trials rising to a maximum of 1. (E) Same as D, only for a simulation with a highly bored population (boredom function of C, ni = 200 individuals, memory size m = 12). Here, individuals show a greater mean trial-to-trial change in color trends over time, accompanied by a lower maximal uniformity. (F) Matrix of simulations across the space of different combinations of the two model parameters, memory size and boredom parameter x0. The color reflects the maximal uniformity of the simulation (simulations with uniformity values between 0.345 and 0.700, representing well-developed collective trends are outlined in white). (G) Absolute speed of change in trend colors (mean angle between subsequent trend color vectors in units of pi) over the parameter space for the same simulations as in F.
