Efficient D2D content caching using multi-agent reinforcement learning | IEEE Conference Publication | IEEE Xplore