Seminal paper on a model of opinion dynamics.
Consider a group of k individuals who must act together as a team. There is some parameter, x. Each individual i has their own subjective probability distribution of the correct value of x, F_i.
Updating
Each individual i assigns a weight p_ij to the opinion of individual j, such that all of i's weights sum to one (and the weights are always non-negative).
Each individual (synchronously) updates his opinions such that:
Convergence
The whole population will converge on a value if the limit of everyone's F is the same. This will happen if and only if there is at least one individual for whom all individuals in the population assign a positive weight. More formally (using the matrix language of the paper):
It is also possible, given everyone's weights of influence and initial probability distributions, to precisely calculate the final consensus.
No comments:
Post a Comment