Multi-Objective Learning Approach to Proactive Caching in Wireless Networks
Te Chen, Yang Du, Binhong Dong, Yantao Chen, Cunlin Zhang · IEEE Communications Letters · 2019
Previous works on learning-based caching problems often only concern prefetching popular contents during off-peak traffic hours and service them to the edge at peak periods. In this letter, we investigate the cache strategy design problem with two possibly conflicting objectives in wireless networks when consumer preferences are unknown. We focus on caching for paid contents and model this decision making problem in a combinatorial multi-objective multi-armed bandit (CMO-MAB) perspective, by taking both the aggregated offloading traffic in dominant objective and the cumulative profit in non-dominant objective into account. Then, we propose the combinatorial scalarizated multi-objective upper confidence bound (CSMO-UCB) and the combinatorial Pareto multi-objective upper confidence bound (CPMO-UCB) algorithms, respectively, to jointly optimize total reward vectors in both objectives. Simulation results demonstrate that the proposed algorithms outperform their competitors, which are not specifically designed to address the caching problems involving dominant and non-dominant objectives.