Modelling Norm Change Through Trendsetters


Şenol B. N., Ova S., HAFIZOĞLU F. M.

22nd European Conference on Multi-Agent Systems, EUMAS 2025, Bucharest, Romanya, 3 - 05 Eylül 2025, cilt.16259 LNCS, ss.308-316, (Tam Metin Bildiri)

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası: 16259 LNCS
  • Doi Numarası: 10.1007/978-3-032-22820-8_21
  • Basıldığı Şehir: Bucharest
  • Basıldığı Ülke: Romanya
  • Sayfa Sayıları: ss.308-316
  • Anahtar Kelimeler: norm change, social norm, trendsetter
  • İstanbul Ticaret Üniversitesi Adresli: Evet

Özet

It is crucial to address the dynamic aspects of norm life cycle in multi-agent systems research. This study presents an agent based model for norm change using trendsetter agents. The agents are connected based on a certain topology, and repeatedly play the coordination game with their immediate neighbors. The agents select a utility maximizing action (along with exploration) based on their past experiences (Q-learning) and social observations (action preferences of nearest neighbors). In order to investigate under which conditions norm change happens, we conducted a comprehensive set of experiments for different trendsetter configurations (position and number), agent related factors (decision model, exploration rate), and topologies. The empirical findings suggest that norm change is sensitive to the trendsetter positions in the network, exploration rates, and the proportion of observed population. A positive relationship is observed between norm change speed and the effect of social observations and the proportion of trendsetters.