Hamta Sedghani

Hamta Sedghani, Zahra Seyedi, Mauro Passacantando et al.

The rapid growth of mobile social networks (MSNs) has significantly increased the demand for low-latency and reliable content delivery, motivating the deployment of edge caching systems. In practice, multiple content providers (CPs) compete for the limited storage resources of edge devices (EDs), while facing heterogeneous budgets and operational costs. This paper investigates a decentralized multi-CP edge caching framework that jointly accounts for CP budget constraints, ED storage limitations, and strategic interactions among all entities. We formulate the interaction between CPs and EDs as a hierarchical game, combining a Stackelberg model for CP-ED interactions with a non-cooperative game among competing CPs. Under light storage constraints, we show that CP competition constitutes an exact potential game, ensuring the existence of a pure-strategy Nash equilibrium and enabling decentralized convergence. When storage constraints are binding, the resulting game loses this structure; nevertheless, extensive simulations demonstrate stable and efficient convergence in practice. Through a comprehensive numerical evaluation, we show that convergence behavior is primarily driven by CP competition rather than the scale of edge infrastructure. We further reveal that storage scarcity fundamentally alters economic outcomes, amplifying inequality among CPs while increasing the relative bargaining power of EDs. The proposed framework provides a scalable and economically grounded solution for decentralized resource allocation in multi-provider edge caching systems.

Hamta Sedghani, Abednego Wamuhindo Kambale, Federica Filippini et al.

Extended reality technologies are transforming fields such as healthcare, entertainment, and education, with Smart Eye-Wears (SEWs) and Artificial Intelligence (AI) playing a crucial role. However, SEWs face inherent limitations in computational power, memory, and battery life, while offloading computations to external servers is constrained by network conditions and server workload variability. To address these challenges, we propose a Federated Reinforcement Learning (FRL) framework, enabling multiple agents to train collaboratively while preserving data privacy. We implemented synchronous and asynchronous federation strategies, where models are aggregated either at fixed intervals or dynamically based on agent progress. Experimental results show that federated agents exhibit significantly lower performance variability, ensuring greater stability and reliability. These findings underscore the potential of FRL for applications requiring robust real-time AI processing, such as real-time object detection in SEWs.