Group formation on a small-world: experiment and modelling

摘要: As a step towards studying human-agent collectives we conduct an online game with human participants cooperating on network. The is presented in the context of achieving group formation through local coordination. players set initially to small world network limited information location other players, coordinate their movements arrange themselves into groups. To understand decision making process construct data-driven model agents based probability matching. allows us gather insight nature and degree rationality employed by players. By varying parameters agent simulations are able benchmark behaviour. We observe that while utilize neighbourhood capacity, perception risk optimal. also find for certain parameter ranges act more efficiently when compared This approach would allow simulate collective dynamics games having strategies playing alongside proxies.