Application of GPU-based IDR(s) Algorithm in Wireless Network Routing Problems

As wireless networks continue to expand, efficiently solving routing optimization problems has become a significant challenge. The sparse matrix equations generated by large-scale network topologies require efficient algorithms during the solution process. Although traditional Krylov subspace methods, such as the Conjugate Gradient method (CG), Generalized Minimal Residual method (GMRES), and Bi-Conjugate Gradient Stabilized method (BiCGSTAB), are capable of handling sparse matrices, they often suffer from slow convergence rates and high computational costs when faced with large-scale networks and complex matrix conditions. This paper proposes a GPU-accelerated IDR(s) algorithm to solve sparse matrix equations in wireless network routing optimization. By introducing multidimensional residuals and a flexible residual management mechanism, IDR(s) effectively reduces the number of iterations. Additionally, the algorithm leverages the parallel computing capabilities of GPUs to enhance computational efficiency. The GPU-based IDR(s) algorithm is applied to routing optimization problems in modern wireless networks.