Patterns in Time: Modelling Schedules in Social Networks

A theoretical account aimed at modeling time usage founded on social network-based multiagent system is proposed. The model introduces mutual, dynamic relationships between the network and schedules: on the one hand, a network is generated by schedules, on the other, schedules are shaped by the network. Time allocation is driven by globally distributed semantic resources. Social network structure and agent time usage have been adjusted by applying evolutionary algorithm to match empirically observed values. The resulting allocation reveals an emergent property which ties the number of agents with activity type priorities. Using the adjusted social network, simulations have been run to investigate time patterns. For that purpose, a time cycle similarity measure has been proposed and computed, unveiling a high amount of regularity, which indirectly supports the view of culture´s function as a latent pattern maintenance.