Intrinsic Price of Anarchy of Age of Information for Converging Sources with Individual Costs

This paper considers a sensor network for real-time monitoring, where multiple sensors can act as equivalent information sources for a common receiver. The sensing goal is to achieve minimal freshness of status updates at the receiver's end, which is captured through the metric known as age of information (AoI). A distributed uncoordinated management is applied to the sensors, so that they send their reports independently and according to a memoryless process. Thus, nodes share the common objective of minimizing AoI but at the same time they incur an individual transmission cost when sending their updates. The goal of the analysis is to evaluate the intrinsic price of anarchy of such a distributed management, even when the nodes are in the best possible conditions, i.e., the information is converging, free from errors and collisions, and fresher updates always pre-empt older ones.