Path-Based Clustering Approaches for a Hybrid Slepian-Wolf Compression and Coded Caching System

Combining caching with source coding, a hybrid content delivery system further facilitates the shift towards Information-Centric Networks. This is a promising technology heralded as the next phase in network design. However, finding the optimal balance between the source coding gains and the computational complexity is itself an NP-hard problem. By modelling the problem using a path-based approach, this paper outlines iterative algorithms that can be tuned to provide control over this trade-off. So too, a necessary condition of optimality is derived. This condition can be applied repeatedly to improve the performance of the results from the iterative algorithms. The Ant Colony Optimisation family of meta-heuristic algorithms is adapted to solve this problem, providing a benchmark that outperforms the Genetic Algorithm presented in prior work. The iterative algorithms have a larger time complexity than other solutions, but still converge in polynomial time. When combined with the optimality condition, they outperform all of the currently proposed algorithms that solve this problem to date. More specifically, this approach produces results that are found to fall in the 99.97th percentile on average.