Deffuant Opinion Dynamics

Bounded confidence model — agents only interact with similar views

The Deffuant model: N agents each hold opinion x∈[0,1]. At each step, a random pair (i,j) is chosen. If |x_i − x_j| < ε (confidence bound), they update: x_i += μ(x_j − x_i), x_j += μ(x_i − x_j). Small ε → fragmentation into many clusters; large ε → consensus. The top panel shows opinion trajectories over time; the bottom panel shows the current opinion histogram.