Online User Scheduling and Resource Allocation for Mobile-Edge Computing Systems
In this paper, we investigate the multi-user mobile edge computing (MEC) system where the number of users in the system is varying. We formulate the multi-user scheduling problem as an infinite horizon Markov decision process (MDP) problem aiming at minimizing the weighted sum of energy consumption and latency of mobile users. For such a large MDP problem, traditional dynamic programming approaches such as value iteration and policy iteration will suffer from the curse of dimensionality. This motivates us to resort to approximate MDP where the key lies at value function approximation. We propose a novel value function approximation approach where the value function is approximated by running a well-designed heuristic baseline policy. Based on the approximated value function, one-step policy iteration is applied to jointly optimize the offloading decision, user selection and transmission power for the selected user. It is theoretically shown that our proposed algorithm has performance improvement compared with the baseline policy. Also, simulation results demonstrate our proposed scheme significantly outperforms the baseline policy and several other heuristic policies.
READ FULL TEXT