Self-Adaptation of Coordination in Imperfectly Known Task Environments

The coordination of multiple agents which are collectively responsible for accomplishing a system's task is among the fundamental issues of organizational design. This paper studies which mode of coordination emerges in systems when the task complexity is imperfectly known, for example, as the system has gone through an external shock. For this, an agent-based simulation based on NK fitness landscapes is employed. The simulation model controls for different levels of task complexity and different degrees of knowledge about the task. The results provide broad support for the conjecture that intense coordination emerges when the complexity of the task to be accomplished collaboratively is high. Moreover, the results suggest that the precision of knowledge about the task subtly may affect which coordination mode emerges.