Cooperative Risk-Sensitive Survivor-Hub Planning for Resilient Multi-Operator RANs

Power outages can disrupt dense urban radio access networks (RANs), particularly when several nearby base-station (BS) sites become unavailable simultaneously. Since equipping every site with long-duration backup energy is costly, this paper investigates the selection of battery-backed survivor hubs in a shared multi-operator RAN. These hardened sites remain operational during outages and recover traffic from failed neigh- boring sites, including those owned by other operators. We formulate hub selection, battery sizing, and scenario-dependent traffic recovery as a two-stage stochastic mixed-integer linear program. The model uses a Conditional Value-at-Risk objective at the 95% confidence level (CVaR0.95), augmented by site- relevance, operator-balance, and recovery-oriented penalties, to limit priority-weighted service loss in severe outage scenarios. Radio-capacity and battery-energy constraints account for the combined native and migrated traffic carried by each hub, while outage duration determines the energy required to maintain operation. We evaluate the framework in a fixed-seed planning case study based on real multi-operator deployment records from Turin, Italy, combined with synthetic traffic and spatially correlated outage scenarios. At the 500 ke reference budget, the proposed method achieves lower empirical tail loss than No Cooperation and the other evaluated reference methods. It is also the only evaluated method whose aggregate demand-weighted recovery exceeds the configured thresholds for all three traffic classes. These results support cooperative survivor-hub planning under the stated case-study assumptions, without constituting an operational performance forecast.