Latency Constrained Partial Offloading and Subcarrier Allocations in Small Cell Networks

Mobile edge computing enables mobile devices to offload their computation intensive applications to servers deployed at the network edge for improving energy efficiency. Multiple mobile users may compete for the wireless network resources when uploading their applications, and coordinating the subcarrier resource allocations among the users is important to optimize the system level energy performance. Partitioning an application further provides the flexibility that allows the mobile device to offload the optimum portion of the application in order to take the best advantage of the available radio resources in computation offloading. In this paper, we study joint partitioning decisions and subcarrier assignments in a multi-cell networks. The objective is to minimize the total energy consumption of the mobile users while satisfying the hard completion time requirements of the applications. The problem is formulated as a mixed-integer nonlinear programming, which is then decomposed into two coupled sub-problems, one linear programming for the offloading decisions under latency constraint of each application, and another matching game with externalities for subcarrier assignments. A joint subcarrier allocation and offloading decision algorithm is proposed, in which matching theory is used to design the user-subcarrier assignments, based on which the optimal offloading ratio is derived for each user. Simulation results demonstrate that the proposed algorithm can greatly reduce the average energy consumption of the mobile users, compared to several other subcarrier assignment and offloading schemes.