From Optimization to Lookup: A Codebook-Based Dynamic User Association Mechanism

Abstract

In terahertz communications, aiding elements (AEs) such as reconfigurable intelligent surfaces and relays can enhance connectivity by establishing alternative paths. However, massive user access and dynamic network conditions make user association highly complex and time-varying, imposing significant real time computational burden. Moreover, the lightweight method that minimizes the AE–user distance may still cause large cascaded path loss due to multiplicative fading. To overcome these challenges, we propose an AE association strategy that restricts the candidate AEs to a specific region and transforms the com putationally intensive real-time optimization of regional shape parameters into an efficient codebook-based search process. Specifically, we first derive the association distance distribution and achievable rate under the proposed strategy. Secondly, the optimal shape parameters are precomputed under different network conditions by solving the achievable rate maximization problem with a metaheuristic algorithm. Meanwhile, the network conditions and their corresponding optimal shape parameters are stored pairwise in a codebook. Under dynamic network conditions, rather than performing computationally intensive re optimization, shape parameters are efficiently retrieved from the codebook using a lightweight lookup to maximize the achievable rate under the current network conditions. Simulation results demonstrate that the region-constrained association strategy can further enhance the achievable rate, and the offline codebook with online lookup mechanism enables efficient determination of the shape parameters under dynamic network conditions. Moreover, the proposed strategy also reduced the computational requirements of edge nodes and lowered the hardware costs.