Caterina De Bacco, Daniel B. Larremore, Cristopher Moore published a new algorithm with the title A physical model for efficient ranking in networks. The model is based on binary interactions among the entities. As often in physical models, interactions via edges are considered as mechanical springs, and the optimal rankings of the nodes are minimizes the total energy (or “energy”) of the system. They show some examples for identifying prestige, dominance, and social hierarchies in human and animal communities.
Further studies will tell how efficient is the new algorithm.
3 thoughts on “A new model for efficient ranking in networks”
This reminds me of multidmensional scaling and multiple multidimensional scaling (Tusnády, Telegdi etc.). They rank things which are characterized by vectorial quantities – by trying to find a curve along which the points representing the objects can be ordered. Somehow graphs had to be embedded and an optimal stucture was found via thinking that the vertices are connected by springs. (This is not a text to be published in Annals of Mathematics :))
Actually, DANIEL LARREMOR will give a lecture on Thursday, https://lsa.umich.edu/cscs/news-events/all-events.detail.html/46448-10489785.html . While occasionally I visit this lecture series, I can’t make this Thursday