Twin Sorting Dynamic Programming Assisted User Association and Wireless Bandwidth Allocation for Hierarchical Federated Learning
This work addresses efficiency improvements for hierarchical federated learning systems, which is incremental as it builds on existing federated learning frameworks with specific optimizations.
The paper tackles the problem of minimizing global round length in hierarchical federated learning by optimizing user association and wireless bandwidth allocation, achieving globally optimal solutions in polynomial time for two edge servers and outperforming alternative schemes in simulations.
In this paper, we study user association and wireless bandwidth allocation for a hierarchical federated learning system that consists of mobile users, edge servers, and a cloud server. To minimize the length of a global round in hierarchical federated learning with equal bandwidth allocation, we formulate a combinatorial optimization problem. We design the twin sorting dynamic programming (TSDP) algorithm that obtains a globally optimal solution in polynomial time when there are two edge servers. In addition, we put forward the TSDP-assisted algorithm for user association when there are three or more edge servers. Furthermore, given a user association matrix, we formulate and solve a convex optimization problem for optimal wireless bandwidth allocation. Simulation results show that the proposed approach outperforms a number of alternative schemes.