TAME: An Efficient Task Allocation Algorithm for Integrated Mobile Gaming

Abstract

We consider an integrated mobile gaming platform, in which the mobile device (e.g., smartphone) of a player can offload some game tasks toward a server as well as some neighboring mobile devices. The advantages of such a platform are manyfold: it can lead to an improved game experience, to a better use of energy resources, and, while offloading tasks to other mobile users, to the exploitation of the unused computing and storage resources of the mobile equipment, thus reducing the bandwidth and computing costs of the overall system. In this context, we formulate the problem of offloading the game computational tasks as an optimization problem that minimizes the maximum energy consumption across a set of mobile devices, under the constraints of a maximum response time and a limited availability of computation, communication, and storage resources. In light of the problem complexity, we then propose a heuristic, called Task Allocation with Minimal Energy cost (TAME), which is shown to closely approximate the optimal solution in all scenarios we considered. TAME also outperforms state-of-the-art algorithms under both synthetic and real scenarios, which have been devised based on a realistic and detailed energy consumption model for computation and communication resources. Our results, although tailored to mobile gaming, could be extended to other applications, where it may be beneficial to offload computational and storage tasks through device-to-device communications, as enabled by Wi-Fi, Bluetooth, or the upcoming 5G technology.