Joint Cache and Computation Resource Optimization in Fog Radio Access Networks

Abstract

Cache resource allocation and computation offloading are envisioned as two promising solutions to reduce maximal delay in fog radio access networks (F-RANs). Inspired by this, a joint cache and computation resource optimization problem is investigated in this paper. We propose a genetic algorithm-based approach that encodes the optimization order rather than the decision into genes and optimizes the delay stepwise. Furthermore, considering that the running time of the algorithm increases rapidly with the size of the network, a strategy with low time complexity is demanded. To achieve distributed cooperation among agents and reduce the complexity, we innovatively design a multi-agent game-based algorithm so that the original problem is decomposed into a priority calculation problem and several distributed sub-problems. Based on calculated priorities, players adjust their decisions in turn to improve their own delay until no player changes actions. Via simulation, the effectiveness of two proposed algorithms in reducing the delay is verified and the game-based algorithm shows lower time complexity and competitive performance. In addition, the impact of nodes’ storage and computation capability on delay performance is demonstrated and analyzed.