Information Bulletin Strategy in Impatient Queuing

Abstract: In Sixth Generation (6G) networks, decentralized control in multi-tenant systems is a suggested enabler for autonomous network operations. However, autonomy requires independent rationale decisions be taken by tenants. This rationality can only be underpinned by timely and continuous access to status information. Despite its importance, the questions of what information should be shared, how much should be communicated, and how frequently updates should be dispatched remain open research challenges. This manuscript proposes an information bulletin strategy defined around two models of the system descriptor states to address these fundamental questions. The strategy is that queues periodically broadcast these information models to tenants at different time intervals, who may respond by reneging from the queue or jockeying to a more favorable one. The expectation is that over time, the queues adapt their processing rates based on what they learn from the tenant behavior. The objective is to minimize overall delay and the impatience. We formulate for this impatience as an optimization problem, whose analytical solution is intractable. We perform numerical experiments to evaluate the performance of the learned queue policy and to assess how closely it approaches optimal conditions.