摘要
The peak age-of-information (AoI) performance in irregular repetition slotted ALOHA (IRSA) is revisited in this paper with a particular focus on heterogeneous networks, where types of users may have different numbers of users, active probabilities, and sensitivities to AoI. A distributed scheduling framework that is both conceptually and practically simple is proposed. A series of analyses and approximations are conducted, which lead to a closed-form approximation of the best scheduling policy within the proposed framework. Extensive simulations are provided to confirm that in spite of being extremely low complex, the proposed policy can significantly reduce the weighted average peak AoI, especially for heavily overloaded scenarios.