Simulating Tolerance in Dynamic Social Networks

This paper studies the concept of tolerance in dynamic social networks where agents are able to make and break connections with neighbors to improve their payoffs. This problem was initially introduced to the authors by observing resistance or tolerance in experiments run in dynamic networks under the two rules that they have developed: the Highest Rewarding Neighborhood rule and the Highest Weighted Reward rule. These rules help agents evaluate their neighbors and decide whether to break a connection or not. They introduce the idea of tolerance in dynamic networks by allowing an agent to maintain a relationship with a bad neighbor for some time. In this research, the authors investigate and define the phenomenon of tolerance in dynamic social networks, particularly with the two rules. The paper defines a mathematical model to predict an agent's tolerance of a bad neighbor and determine the factors that affect it. After defining a general version of tolerance, the idea of optimal tolerance is explored, providing situations in which tolerance can be used as a tool to affect network efficiency and network structure.