Bounded Confidence Opinion Dynamics
Deffuant model: cluster formation & polarization
Model Parameters
Model:
Deffuant (pairwise)
Hegselmann-Krause (all)
Agents N:
100
Confidence ε:
0.25
Convergence μ:
0.50
Extremists fraction:
0.05
Reset / Randomize
Run 100 steps
Auto-run
Deffuant:
At each step, pick random pair (i,j). If |x_i - x_j| < ε: x_i += μ(x_j-x_i), x_j += μ(x_i-x_j)
HK:
All agents update simultaneously towards local average of neighbors within ε.
Critical ε:
ε ≈ 0.5: consensus
0.3 < ε < 0.5: 2 clusters
ε < 0.3: polarization