On Dynamic Node Cooperation Strategy Design for Energy Efficiency in Hierarchical Federated Learning

In Hierarchical Federated Learning (HFL), opportunistic communication provides opportunities for node cooperation. In this work, we optimize the node cooperation strategy using opportunistic communization with the objective to minimize energy cost under the delay constraint. We design an online node cooperation strategy (OSRN) based on the optimal stopping theory. Through theoretical analysis, we prove the NP-hardness of the problem investigated and the competition ratio that can be achieved by OSRN. We conduct thorough simulation experiments and find that the proposed algorithm outperforms the random selection algorithm SNNR with 22.04% reduction in energy cost. It is also observed that the energy cost can be reduced by 20.20% and 13.54%, respectively, compared with the existing methods CFL and THF.